MNIST-Traditional Filters.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import tensorflow.keras as keras\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 1. Load the prepared dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cifar10', 'mnist']"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "DATA_DIR = 'C:/Users/MINH LE/Desktop/TraditionalvsCNN/dataset/'\n",
    "\n",
    "list_files = os.listdir(DATA_DIR)\n",
    "number_files = len(list_files)\n",
    "list_files"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.1. MNIST Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CLTP FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import glob, cv2\n",
    "cltp = []\n",
    "\n",
    "CLTP_DIR = DATA_DIR + 'mnist/cltp/train/'\n",
    "for img in glob.glob(CLTP_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    cltp.append(n)\n",
    "    \n",
    "cltp = np.asarray(cltp, dtype=np.float32)\n",
    "cltp = np.reshape(cltp, (60000, 26, 26, 1))\n",
    "cltp.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**CSLBP FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cslbp = []\n",
    "\n",
    "CSLBP_DIR = DATA_DIR + 'mnist/cslbp/train/'\n",
    "for img in glob.glob(CSLBP_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    cslbp.append(n)\n",
    "    \n",
    "cslbp = np.asarray(cslbp, dtype=np.float32)\n",
    "cslbp = np.reshape(cslbp, (60000, 26, 26, 1))\n",
    "cslbp.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**GAUSSIAN FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gauss_filter = []\n",
    "\n",
    "GAUSSIAN_DIR = DATA_DIR + 'mnist/gaussian/train/'\n",
    "for img in glob.glob(GAUSSIAN_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    gauss_filter.append(n)\n",
    "    \n",
    "gauss_filter = np.asarray(gauss_filter, dtype=np.float32)\n",
    "gauss_filter = np.reshape(gauss_filter, (60000, 26, 26, 1))\n",
    "gauss_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**GLTP FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gltp_filter = []\n",
    "\n",
    "GLTP_DIR = DATA_DIR + 'mnist/gltp/train/'\n",
    "for img in glob.glob(GLTP_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    gltp_filter.append(n)\n",
    "    \n",
    "gltp_filter = np.asarray(gltp_filter, dtype=np.float32)\n",
    "gltp_filter = np.reshape(gltp_filter, (60000, 26, 26, 1))\n",
    "gltp_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**LAPLACIAN FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "laplacian_filter = []\n",
    "\n",
    "LAPLACIAN_DIR = DATA_DIR + 'mnist/laplacian/train/'\n",
    "for img in glob.glob(LAPLACIAN_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    laplacian_filter.append(n)\n",
    "    \n",
    "laplacian_filter = np.asarray(laplacian_filter, dtype=np.float32)\n",
    "laplacian_filter = np.reshape(laplacian_filter, (60000, 26, 26, 1))\n",
    "laplacian_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**LOCAL BINARY PATTERN FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lbp_filter = []\n",
    "\n",
    "LBP_DIR = DATA_DIR + 'mnist/lbp/train/'\n",
    "for img in glob.glob(LBP_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    lbp_filter.append(n)\n",
    "    \n",
    "lbp_filter = np.asarray(lbp_filter, dtype=np.float32)\n",
    "lbp_filter = np.reshape(lbp_filter, (60000, 26, 26, 1))\n",
    "lbp_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**LOCAL TERNARY PATTERN**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ltp_filter = []\n",
    "\n",
    "LTP_DIR = DATA_DIR + 'mnist/ltp/train/'\n",
    "for img in glob.glob(LTP_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    ltp_filter.append(n)\n",
    "    \n",
    "ltp_filter = np.asarray(ltp_filter, dtype=np.float32)\n",
    "ltp_filter = np.reshape(ltp_filter, (60000, 26, 26, 1))\n",
    "ltp_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**PREWITT X FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prewittx_filter = []\n",
    "\n",
    "PREWITTX_DIR = DATA_DIR + 'mnist/prewittx/train/'\n",
    "for img in glob.glob(PREWITTX_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    prewittx_filter.append(n)\n",
    "    \n",
    "prewittx_filter = np.asarray(prewittx_filter, dtype=np.float32)\n",
    "prewittx_filter = np.reshape(prewittx_filter, (60000, 26, 26, 1))\n",
    "prewittx_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**PREWITT Y FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prewitty_filter = []\n",
    "\n",
    "PREWITTY_DIR = DATA_DIR + 'mnist/prewitty/train/'\n",
    "for img in glob.glob(PREWITTY_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    prewitty_filter.append(n)\n",
    "    \n",
    "prewitty_filter = np.asarray(prewitty_filter, dtype=np.float32)\n",
    "prewitty_filter = np.reshape(prewitty_filter, (60000, 26, 26, 1))\n",
    "prewitty_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**SCS-LTP FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scsltp_filter = []\n",
    "\n",
    "SCSLTP_DIR = DATA_DIR + 'mnist/scsltp/train/'\n",
    "for img in glob.glob(SCSLTP_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    scsltp_filter.append(n)\n",
    "    \n",
    "scsltp_filter = np.asarray(scsltp_filter, dtype=np.float32)\n",
    "scsltp_filter = np.reshape(scsltp_filter, (60000, 26, 26, 1))\n",
    "scsltp_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**SILTP FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "siltp_filter = []\n",
    "\n",
    "SILTP_DIR = DATA_DIR + 'mnist/siltp/train/'\n",
    "for img in glob.glob(SILTP_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    siltp_filter.append(n)\n",
    "    \n",
    "siltp_filter = np.asarray(siltp_filter, dtype=np.float32)\n",
    "siltp_filter = np.reshape(siltp_filter, (60000, 26, 26, 1))\n",
    "siltp_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**SOBEL X FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sobelx_filter = []\n",
    "\n",
    "SOBELX_DIR = DATA_DIR + 'mnist/sobelx/train/'\n",
    "for img in glob.glob(SOBELX_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    sobelx_filter.append(n)\n",
    "    \n",
    "sobelx_filter = np.asarray(sobelx_filter, dtype=np.float32)\n",
    "sobelx_filter = np.reshape(sobelx_filter, (60000, 26, 26, 1))\n",
    "sobelx_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**SOBEL Y FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sobely_filter = []\n",
    "\n",
    "SOBELY_DIR = DATA_DIR + 'mnist/sobely/train/'\n",
    "for img in glob.glob(SOBELY_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    sobely_filter.append(n)\n",
    "    \n",
    "sobely_filter = np.asarray(sobely_filter, dtype=np.float32)\n",
    "sobely_filter = np.reshape(sobely_filter, (60000, 26, 26, 1))\n",
    "sobely_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**XCSLTP FILTER**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 26, 26, 1)"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xcsltp_filter = []\n",
    "\n",
    "XCSLTP_DIR = DATA_DIR + 'mnist/xcsltp/train/'\n",
    "for img in glob.glob(XCSLTP_DIR + '*.jpg'):\n",
    "    n = cv2.imread(img, 0)\n",
    "    xcsltp_filter.append(n)\n",
    "    \n",
    "xcsltp_filter = np.asarray(xcsltp_filter, dtype=np.float32)\n",
    "xcsltp_filter = np.reshape(xcsltp_filter, (60000, 26, 26, 1))\n",
    "xcsltp_filter.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2. Training Phase"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.layers import Flatten\n",
    "from tensorflow.keras.layers import MaxPooling2D\n",
    "from tensorflow.keras.layers import Input\n",
    "from tensorflow.keras.layers import Dropout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([60000, 26, 26, 14])"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data = tf.keras.layers.Concatenate()([cltp, \n",
    "                                            cslbp, \n",
    "                                            gauss_filter, \n",
    "                                            gltp_filter,\n",
    "                                            laplacian_filter,\n",
    "                                            lbp_filter,\n",
    "                                            ltp_filter,\n",
    "                                            prewittx_filter,\n",
    "                                            prewitty_filter,\n",
    "                                            scsltp_filter,\n",
    "                                            siltp_filter,\n",
    "                                            sobelx_filter,\n",
    "                                            sobely_filter,\n",
    "                                            xcsltp_filter])\n",
    "train_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "TensorShape([60000, 26, 26, 14])"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_data = tf.reshape(train_data, shape=(60000, 26, 26, 14))\n",
    "\n",
    "train_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "mnist = keras.datasets.mnist\n",
    "(_, train_labels), (_, _) = mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: keras in c:\\users\\minh le\\anaconda3\\lib\\site-packages (2.4.3)\n",
      "Requirement already satisfied: numpy>=1.9.1 in c:\\users\\minh le\\anaconda3\\lib\\site-packages (from keras) (1.19.5)\n",
      "Requirement already satisfied: h5py in c:\\users\\minh le\\anaconda3\\lib\\site-packages (from keras) (2.10.0)\n",
      "Requirement already satisfied: pyyaml in c:\\users\\minh le\\anaconda3\\lib\\site-packages (from keras) (5.3.1)\n",
      "Requirement already satisfied: scipy>=0.14 in c:\\users\\minh le\\anaconda3\\lib\\site-packages (from keras) (1.6.0)\n",
      "Requirement already satisfied: six in c:\\users\\minh le\\anaconda3\\lib\\site-packages (from h5py->keras) (1.15.0)\n"
     ]
    }
   ],
   "source": [
    "!pip install keras"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Experiment 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/300\n",
      "59/59 [==============================] - 6s 87ms/step - loss: 978.4411 - accuracy: 0.0991\n",
      "Epoch 2/300\n",
      "59/59 [==============================] - 4s 73ms/step - loss: 13.5631 - accuracy: 0.1014\n",
      "Epoch 3/300\n",
      "59/59 [==============================] - 4s 71ms/step - loss: 6.7182 - accuracy: 0.1019\n",
      "Epoch 4/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 4.5178 - accuracy: 0.1017\n",
      "Epoch 5/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 3.2890 - accuracy: 0.0993\n",
      "Epoch 6/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3044 - accuracy: 0.1042 0s - loss: 2.3044 - accuracy: 0.\n",
      "Epoch 7/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3037 - accuracy: 0.1134\n",
      "Epoch 8/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3023 - accuracy: 0.1128\n",
      "Epoch 9/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3023 - accuracy: 0.1119\n",
      "Epoch 10/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3023 - accuracy: 0.1117\n",
      "Epoch 11/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3015 - accuracy: 0.1124\n",
      "Epoch 12/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3008 - accuracy: 0.1132\n",
      "Epoch 13/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3008 - accuracy: 0.1133\n",
      "Epoch 14/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3010 - accuracy: 0.1117\n",
      "Epoch 15/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3003 - accuracy: 0.1134\n",
      "Epoch 16/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3005 - accuracy: 0.1135\n",
      "Epoch 17/300\n",
      "59/59 [==============================] - 4s 59ms/step - loss: 2.3005 - accuracy: 0.1124\n",
      "Epoch 18/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3001 - accuracy: 0.1135\n",
      "Epoch 19/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3002 - accuracy: 0.1123\n",
      "Epoch 20/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3000 - accuracy: 0.1125\n",
      "Epoch 21/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3001 - accuracy: 0.1127\n",
      "Epoch 22/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2995 - accuracy: 0.1142\n",
      "Epoch 23/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2995 - accuracy: 0.1126\n",
      "Epoch 24/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2998 - accuracy: 0.1137\n",
      "Epoch 25/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2995 - accuracy: 0.1121\n",
      "Epoch 26/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2998 - accuracy: 0.1120\n",
      "Epoch 27/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2999 - accuracy: 0.1115\n",
      "Epoch 28/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2987 - accuracy: 0.1149\n",
      "Epoch 29/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2989 - accuracy: 0.1154\n",
      "Epoch 30/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2994 - accuracy: 0.1110\n",
      "Epoch 31/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2995 - accuracy: 0.1128\n",
      "Epoch 32/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2989 - accuracy: 0.1135\n",
      "Epoch 33/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2991 - accuracy: 0.1144\n",
      "Epoch 34/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2993 - accuracy: 0.1125\n",
      "Epoch 35/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2990 - accuracy: 0.1141\n",
      "Epoch 36/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2992 - accuracy: 0.1137\n",
      "Epoch 37/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2990 - accuracy: 0.1139\n",
      "Epoch 38/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2991 - accuracy: 0.1126\n",
      "Epoch 39/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2991 - accuracy: 0.1120\n",
      "Epoch 40/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2986 - accuracy: 0.1139\n",
      "Epoch 41/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2986 - accuracy: 0.1143\n",
      "Epoch 42/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2989 - accuracy: 0.1136\n",
      "Epoch 43/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.2991 - accuracy: 0.1124\n",
      "Epoch 44/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2995 - accuracy: 0.1117\n",
      "Epoch 45/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2988 - accuracy: 0.1135\n",
      "Epoch 46/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2987 - accuracy: 0.1132\n",
      "Epoch 47/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2993 - accuracy: 0.1121\n",
      "Epoch 48/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2992 - accuracy: 0.1113\n",
      "Epoch 49/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2986 - accuracy: 0.1141\n",
      "Epoch 50/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2987 - accuracy: 0.1128\n",
      "Epoch 51/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2987 - accuracy: 0.1122\n",
      "Epoch 52/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2985 - accuracy: 0.1135\n",
      "Epoch 53/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2981 - accuracy: 0.1149\n",
      "Epoch 54/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2984 - accuracy: 0.1129\n",
      "Epoch 55/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2986 - accuracy: 0.1121\n",
      "Epoch 56/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.2985 - accuracy: 0.1118\n",
      "Epoch 57/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.2982 - accuracy: 0.1141\n",
      "Epoch 58/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.2983 - accuracy: 0.1117\n",
      "Epoch 59/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.2979 - accuracy: 0.1132\n",
      "Epoch 60/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.2985 - accuracy: 0.1105\n",
      "Epoch 61/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.2980 - accuracy: 0.1120\n",
      "Epoch 62/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.2983 - accuracy: 0.1111\n",
      "Epoch 63/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.2982 - accuracy: 0.1141\n",
      "Epoch 64/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.2980 - accuracy: 0.1122\n",
      "Epoch 65/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.2981 - accuracy: 0.1128\n",
      "Epoch 66/300\n",
      "59/59 [==============================] - 4s 74ms/step - loss: 2.2980 - accuracy: 0.1130\n",
      "Epoch 67/300\n",
      "59/59 [==============================] - 4s 71ms/step - loss: 2.2982 - accuracy: 0.1115\n",
      "Epoch 68/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.2975 - accuracy: 0.1130\n",
      "Epoch 69/300\n",
      "59/59 [==============================] - 4s 73ms/step - loss: 2.2973 - accuracy: 0.1150\n",
      "Epoch 70/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.2979 - accuracy: 0.1123\n",
      "Epoch 71/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.2972 - accuracy: 0.1128\n",
      "Epoch 72/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2972 - accuracy: 0.1154\n",
      "Epoch 73/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2969 - accuracy: 0.1137\n",
      "Epoch 74/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2973 - accuracy: 0.1140\n",
      "Epoch 75/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.2977 - accuracy: 0.1140\n",
      "Epoch 76/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2972 - accuracy: 0.1142\n",
      "Epoch 77/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2976 - accuracy: 0.1124\n",
      "Epoch 78/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2964 - accuracy: 0.1151\n",
      "Epoch 79/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2975 - accuracy: 0.1120\n",
      "Epoch 80/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.2975 - accuracy: 0.1112\n",
      "Epoch 81/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2973 - accuracy: 0.1124\n",
      "Epoch 82/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2981 - accuracy: 0.1108\n",
      "Epoch 83/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2976 - accuracy: 0.1139\n",
      "Epoch 84/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2970 - accuracy: 0.1117\n",
      "Epoch 85/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2974 - accuracy: 0.1134\n",
      "Epoch 86/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2963 - accuracy: 0.1137\n",
      "Epoch 87/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2963 - accuracy: 0.1134\n",
      "Epoch 88/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2974 - accuracy: 0.1125\n",
      "Epoch 89/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2966 - accuracy: 0.1143\n",
      "Epoch 90/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2962 - accuracy: 0.1144\n",
      "Epoch 91/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2958 - accuracy: 0.1143\n",
      "Epoch 92/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2962 - accuracy: 0.1134\n",
      "Epoch 93/300\n",
      "59/59 [==============================] - 3s 59ms/step - loss: 2.2962 - accuracy: 0.1134\n",
      "Epoch 94/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2964 - accuracy: 0.1144\n",
      "Epoch 95/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2967 - accuracy: 0.1138\n",
      "Epoch 96/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2962 - accuracy: 0.1128\n",
      "Epoch 97/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2959 - accuracy: 0.1138\n",
      "Epoch 98/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2958 - accuracy: 0.1132\n",
      "Epoch 99/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2969 - accuracy: 0.1112\n",
      "Epoch 100/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2971 - accuracy: 0.1136\n",
      "Epoch 101/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.2978 - accuracy: 0.1133\n",
      "Epoch 102/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2979 - accuracy: 0.1129\n",
      "Epoch 103/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2981 - accuracy: 0.1124\n",
      "Epoch 104/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2981 - accuracy: 0.1128\n",
      "Epoch 105/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2985 - accuracy: 0.1115\n",
      "Epoch 106/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.2981 - accuracy: 0.1131\n",
      "Epoch 107/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2978 - accuracy: 0.1149\n",
      "Epoch 108/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2989 - accuracy: 0.1131\n",
      "Epoch 109/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2983 - accuracy: 0.1137\n",
      "Epoch 110/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.2989 - accuracy: 0.1111\n",
      "Epoch 111/300\n",
      "59/59 [==============================] - 5s 76ms/step - loss: 2.2976 - accuracy: 0.1142\n",
      "Epoch 112/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2983 - accuracy: 0.1127\n",
      "Epoch 113/300\n",
      "59/59 [==============================] - 5s 77ms/step - loss: 2.2969 - accuracy: 0.1154\n",
      "Epoch 114/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.2981 - accuracy: 0.1127\n",
      "Epoch 115/300\n",
      "59/59 [==============================] - 5s 81ms/step - loss: 2.2993 - accuracy: 0.1145\n",
      "Epoch 116/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.2984 - accuracy: 0.1128\n",
      "Epoch 117/300\n",
      "59/59 [==============================] - 4s 72ms/step - loss: 2.2988 - accuracy: 0.1120\n",
      "Epoch 118/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2994 - accuracy: 0.1145\n",
      "Epoch 119/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2986 - accuracy: 0.1130\n",
      "Epoch 120/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2985 - accuracy: 0.1118\n",
      "Epoch 121/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2986 - accuracy: 0.1127\n",
      "Epoch 122/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2986 - accuracy: 0.1122\n",
      "Epoch 123/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2992 - accuracy: 0.1131\n",
      "Epoch 124/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.2989 - accuracy: 0.1129\n",
      "Epoch 125/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.2995 - accuracy: 0.1121\n",
      "Epoch 126/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2992 - accuracy: 0.1134\n",
      "Epoch 127/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2991 - accuracy: 0.1129\n",
      "Epoch 128/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2989 - accuracy: 0.1123\n",
      "Epoch 129/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.2996 - accuracy: 0.1119\n",
      "Epoch 130/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2993 - accuracy: 0.1150\n",
      "Epoch 131/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2996 - accuracy: 0.1125\n",
      "Epoch 132/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.2995 - accuracy: 0.1120\n",
      "Epoch 133/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.2991 - accuracy: 0.1128\n",
      "Epoch 134/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.2998 - accuracy: 0.1127\n",
      "Epoch 135/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.3001 - accuracy: 0.1127\n",
      "Epoch 136/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.3010 - accuracy: 0.1113\n",
      "Epoch 137/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3013 - accuracy: 0.1120\n",
      "Epoch 138/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3008 - accuracy: 0.1123\n",
      "Epoch 139/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3010 - accuracy: 0.1108\n",
      "Epoch 140/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3009 - accuracy: 0.1128\n",
      "Epoch 141/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.3012 - accuracy: 0.1105\n",
      "Epoch 142/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3013 - accuracy: 0.1113\n",
      "Epoch 143/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3010 - accuracy: 0.1124\n",
      "Epoch 144/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3013 - accuracy: 0.1118\n",
      "Epoch 145/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3011 - accuracy: 0.1124\n",
      "Epoch 146/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3011 - accuracy: 0.1118\n",
      "Epoch 147/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.3010 - accuracy: 0.1134\n",
      "Epoch 148/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.3012 - accuracy: 0.1113\n",
      "Epoch 149/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3010 - accuracy: 0.1121\n",
      "Epoch 150/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.3011 - accuracy: 0.1122\n",
      "Epoch 151/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3011 - accuracy: 0.1142\n",
      "Epoch 152/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.3010 - accuracy: 0.1120\n",
      "Epoch 153/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3012 - accuracy: 0.1120\n",
      "Epoch 154/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3014 - accuracy: 0.1112\n",
      "Epoch 155/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3012 - accuracy: 0.1122\n",
      "Epoch 156/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3007 - accuracy: 0.1131\n",
      "Epoch 157/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3009 - accuracy: 0.1120\n",
      "Epoch 158/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3012 - accuracy: 0.1126\n",
      "Epoch 159/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3011 - accuracy: 0.1116\n",
      "Epoch 160/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3015 - accuracy: 0.1110\n",
      "Epoch 161/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3015 - accuracy: 0.1107\n",
      "Epoch 162/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3011 - accuracy: 0.1111\n",
      "Epoch 163/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3007 - accuracy: 0.1141\n",
      "Epoch 164/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3010 - accuracy: 0.1126\n",
      "Epoch 165/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3010 - accuracy: 0.1125\n",
      "Epoch 166/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3012 - accuracy: 0.1115\n",
      "Epoch 167/300\n",
      "59/59 [==============================] - 4s 59ms/step - loss: 2.3010 - accuracy: 0.1136\n",
      "Epoch 168/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3011 - accuracy: 0.1124\n",
      "Epoch 169/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3010 - accuracy: 0.1116\n",
      "Epoch 170/300\n",
      "59/59 [==============================] - 4s 59ms/step - loss: 2.3010 - accuracy: 0.1133\n",
      "Epoch 171/300\n",
      "59/59 [==============================] - 3s 59ms/step - loss: 2.3014 - accuracy: 0.1120\n",
      "Epoch 172/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3012 - accuracy: 0.1108\n",
      "Epoch 173/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3011 - accuracy: 0.1139\n",
      "Epoch 174/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3012 - accuracy: 0.1129\n",
      "Epoch 175/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3012 - accuracy: 0.1117\n",
      "Epoch 176/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3011 - accuracy: 0.1126\n",
      "Epoch 177/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3013 - accuracy: 0.1109\n",
      "Epoch 178/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3010 - accuracy: 0.1112\n",
      "Epoch 179/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3010 - accuracy: 0.1133\n",
      "Epoch 180/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3012 - accuracy: 0.1117\n",
      "Epoch 181/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3012 - accuracy: 0.1132\n",
      "Epoch 182/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3013 - accuracy: 0.1114\n",
      "Epoch 183/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3013 - accuracy: 0.1125\n",
      "Epoch 184/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3012 - accuracy: 0.1134\n",
      "Epoch 185/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3012 - accuracy: 0.1128\n",
      "Epoch 186/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3009 - accuracy: 0.1146\n",
      "Epoch 187/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3015 - accuracy: 0.1112\n",
      "Epoch 188/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3010 - accuracy: 0.1129\n",
      "Epoch 189/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3014 - accuracy: 0.1116\n",
      "Epoch 190/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3008 - accuracy: 0.1123\n",
      "Epoch 191/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3015 - accuracy: 0.1111\n",
      "Epoch 192/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3010 - accuracy: 0.1125\n",
      "Epoch 193/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3014 - accuracy: 0.1117\n",
      "Epoch 194/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3012 - accuracy: 0.1119\n",
      "Epoch 195/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3013 - accuracy: 0.1114\n",
      "Epoch 196/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3009 - accuracy: 0.1141\n",
      "Epoch 197/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3007 - accuracy: 0.1146\n",
      "Epoch 198/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3012 - accuracy: 0.1126\n",
      "Epoch 199/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3011 - accuracy: 0.1118\n",
      "Epoch 200/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3013 - accuracy: 0.1121\n",
      "Epoch 201/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3013 - accuracy: 0.1125\n",
      "Epoch 202/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3007 - accuracy: 0.1136\n",
      "Epoch 203/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3014 - accuracy: 0.1119\n",
      "Epoch 204/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.3011 - accuracy: 0.1122\n",
      "Epoch 205/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3010 - accuracy: 0.1137\n",
      "Epoch 206/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3007 - accuracy: 0.1132\n",
      "Epoch 207/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3011 - accuracy: 0.1141\n",
      "Epoch 208/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.3013 - accuracy: 0.1123\n",
      "Epoch 209/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.3013 - accuracy: 0.1118\n",
      "Epoch 210/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.3013 - accuracy: 0.1113\n",
      "Epoch 211/300\n",
      "59/59 [==============================] - 4s 71ms/step - loss: 2.3017 - accuracy: 0.1106\n",
      "Epoch 212/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3008 - accuracy: 0.1152\n",
      "Epoch 213/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3014 - accuracy: 0.1103\n",
      "Epoch 214/300\n",
      "59/59 [==============================] - 4s 71ms/step - loss: 2.3011 - accuracy: 0.1129\n",
      "Epoch 215/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3013 - accuracy: 0.1122\n",
      "Epoch 216/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3012 - accuracy: 0.1123\n",
      "Epoch 217/300\n",
      "59/59 [==============================] - 4s 71ms/step - loss: 2.3012 - accuracy: 0.1121\n",
      "Epoch 218/300\n",
      "59/59 [==============================] - 4s 72ms/step - loss: 2.3011 - accuracy: 0.1126\n",
      "Epoch 219/300\n",
      "59/59 [==============================] - 4s 71ms/step - loss: 2.3014 - accuracy: 0.1119\n",
      "Epoch 220/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3010 - accuracy: 0.1133\n",
      "Epoch 221/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3008 - accuracy: 0.1127\n",
      "Epoch 222/300\n",
      "59/59 [==============================] - 4s 71ms/step - loss: 2.3009 - accuracy: 0.1147\n",
      "Epoch 223/300\n",
      "59/59 [==============================] - 4s 75ms/step - loss: 2.3010 - accuracy: 0.1126\n",
      "Epoch 224/300\n",
      "59/59 [==============================] - 4s 74ms/step - loss: 2.3014 - accuracy: 0.1102\n",
      "Epoch 225/300\n",
      "59/59 [==============================] - 4s 76ms/step - loss: 2.3009 - accuracy: 0.1125\n",
      "Epoch 226/300\n",
      "59/59 [==============================] - 4s 73ms/step - loss: 2.3011 - accuracy: 0.1128\n",
      "Epoch 227/300\n",
      "59/59 [==============================] - 4s 72ms/step - loss: 2.3015 - accuracy: 0.1108\n",
      "Epoch 228/300\n",
      "59/59 [==============================] - 4s 72ms/step - loss: 2.3013 - accuracy: 0.1129\n",
      "Epoch 229/300\n",
      "59/59 [==============================] - 4s 71ms/step - loss: 2.3008 - accuracy: 0.1137\n",
      "Epoch 230/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3013 - accuracy: 0.1114\n",
      "Epoch 231/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.3013 - accuracy: 0.1122\n",
      "Epoch 232/300\n",
      "59/59 [==============================] - 4s 69ms/step - loss: 2.3016 - accuracy: 0.1096\n",
      "Epoch 233/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3014 - accuracy: 0.1119\n",
      "Epoch 234/300\n",
      "59/59 [==============================] - 4s 68ms/step - loss: 2.3011 - accuracy: 0.1130\n",
      "Epoch 235/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.3013 - accuracy: 0.1119\n",
      "Epoch 236/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3011 - accuracy: 0.1124\n",
      "Epoch 237/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3010 - accuracy: 0.1123\n",
      "Epoch 238/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3014 - accuracy: 0.1117\n",
      "Epoch 239/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3009 - accuracy: 0.1148\n",
      "Epoch 240/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3010 - accuracy: 0.1147\n",
      "Epoch 241/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3013 - accuracy: 0.1124\n",
      "Epoch 242/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3010 - accuracy: 0.1141\n",
      "Epoch 243/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3012 - accuracy: 0.1128\n",
      "Epoch 244/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3012 - accuracy: 0.1127\n",
      "Epoch 245/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3014 - accuracy: 0.1109\n",
      "Epoch 246/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3009 - accuracy: 0.1134\n",
      "Epoch 247/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3016 - accuracy: 0.1108\n",
      "Epoch 248/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3013 - accuracy: 0.1122\n",
      "Epoch 249/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3010 - accuracy: 0.1143\n",
      "Epoch 250/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3013 - accuracy: 0.1119\n",
      "Epoch 251/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3008 - accuracy: 0.1131\n",
      "Epoch 252/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3013 - accuracy: 0.1108\n",
      "Epoch 253/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3014 - accuracy: 0.1125\n",
      "Epoch 254/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3012 - accuracy: 0.1119\n",
      "Epoch 255/300\n",
      "59/59 [==============================] - 3s 59ms/step - loss: 2.3009 - accuracy: 0.1132\n",
      "Epoch 256/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3009 - accuracy: 0.1132\n",
      "Epoch 257/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3011 - accuracy: 0.1123\n",
      "Epoch 258/300\n",
      "59/59 [==============================] - 3s 59ms/step - loss: 2.3012 - accuracy: 0.1135\n",
      "Epoch 259/300\n",
      "59/59 [==============================] - 3s 59ms/step - loss: 2.3011 - accuracy: 0.1133\n",
      "Epoch 260/300\n",
      "59/59 [==============================] - 3s 59ms/step - loss: 2.3014 - accuracy: 0.1121\n",
      "Epoch 261/300\n",
      "59/59 [==============================] - 3s 59ms/step - loss: 2.3011 - accuracy: 0.1114\n",
      "Epoch 262/300\n",
      "59/59 [==============================] - 3s 59ms/step - loss: 2.3010 - accuracy: 0.1144\n",
      "Epoch 263/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3011 - accuracy: 0.1127\n",
      "Epoch 264/300\n",
      "59/59 [==============================] - 4s 61ms/step - loss: 2.3013 - accuracy: 0.1125\n",
      "Epoch 265/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3012 - accuracy: 0.1123\n",
      "Epoch 266/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3015 - accuracy: 0.1108\n",
      "Epoch 267/300\n",
      "59/59 [==============================] - 4s 60ms/step - loss: 2.3012 - accuracy: 0.1125\n",
      "Epoch 268/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3014 - accuracy: 0.1110\n",
      "Epoch 269/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.3013 - accuracy: 0.1133\n",
      "Epoch 270/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3011 - accuracy: 0.1123\n",
      "Epoch 271/300\n",
      "59/59 [==============================] - 4s 62ms/step - loss: 2.3012 - accuracy: 0.1121\n",
      "Epoch 272/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3008 - accuracy: 0.1140\n",
      "Epoch 273/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3014 - accuracy: 0.1107\n",
      "Epoch 274/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3013 - accuracy: 0.1114\n",
      "Epoch 275/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3013 - accuracy: 0.1107\n",
      "Epoch 276/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3011 - accuracy: 0.1112\n",
      "Epoch 277/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3012 - accuracy: 0.1119\n",
      "Epoch 278/300\n",
      "59/59 [==============================] - 4s 63ms/step - loss: 2.3012 - accuracy: 0.1114\n",
      "Epoch 279/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3012 - accuracy: 0.1145\n",
      "Epoch 280/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3010 - accuracy: 0.1134\n",
      "Epoch 281/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3012 - accuracy: 0.1109\n",
      "Epoch 282/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3013 - accuracy: 0.1102\n",
      "Epoch 283/300\n",
      "59/59 [==============================] - 4s 64ms/step - loss: 2.3014 - accuracy: 0.1110\n",
      "Epoch 284/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3013 - accuracy: 0.1117\n",
      "Epoch 285/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3007 - accuracy: 0.1127\n",
      "Epoch 286/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3007 - accuracy: 0.1143\n",
      "Epoch 287/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3011 - accuracy: 0.1124\n",
      "Epoch 288/300\n",
      "59/59 [==============================] - 4s 65ms/step - loss: 2.3015 - accuracy: 0.1116\n",
      "Epoch 289/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3012 - accuracy: 0.1123\n",
      "Epoch 290/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3010 - accuracy: 0.1118\n",
      "Epoch 291/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3010 - accuracy: 0.1142\n",
      "Epoch 292/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3012 - accuracy: 0.1129\n",
      "Epoch 293/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3014 - accuracy: 0.1134\n",
      "Epoch 294/300\n",
      "59/59 [==============================] - 4s 66ms/step - loss: 2.3017 - accuracy: 0.1117\n",
      "Epoch 295/300\n",
      "59/59 [==============================] - 4s 67ms/step - loss: 2.3014 - accuracy: 0.1092\n",
      "Epoch 296/300\n",
      "59/59 [==============================] - 4s 73ms/step - loss: 2.3015 - accuracy: 0.1104\n",
      "Epoch 297/300\n",
      "59/59 [==============================] - 4s 70ms/step - loss: 2.3011 - accuracy: 0.1138\n",
      "Epoch 298/300\n",
      "59/59 [==============================] - 4s 74ms/step - loss: 2.3012 - accuracy: 0.1133\n",
      "Epoch 299/300\n",
      "59/59 [==============================] - 4s 72ms/step - loss: 2.3012 - accuracy: 0.1136\n",
      "Epoch 300/300\n",
      "59/59 [==============================] - 4s 72ms/step - loss: 2.3009 - accuracy: 0.1145\n",
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "max_pooling2d (MaxPooling2D) (None, 13, 13, 14)        0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 2366)              0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 1024)              2423808   \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 10)                10250     \n",
      "=================================================================\n",
      "Total params: 2,434,058\n",
      "Trainable params: 2,434,058\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "\n",
    "# Initialize the He initializer\n",
    "initializer = tf.keras.initializers.HeNormal()\n",
    "model_1 = Sequential()\n",
    "\n",
    "model_1.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model_1.add(Flatten())\n",
    "model_1.add(Dense(1024, activation='relu', kernel_initializer=initializer))\n",
    "model_1.add(Dense(10, activation='softmax', kernel_initializer=initializer))\n",
    "\n",
    "model_1.compile(optimizer='adam', \n",
    "               loss='sparse_categorical_crossentropy', \n",
    "               metrics = ['accuracy'])\n",
    "\n",
    "history_data1 = model_1.fit(x=train_data, \n",
    "                            y=train_labels, \n",
    "                            epochs=300, \n",
    "                            batch_size=1024)\n",
    "\n",
    "model_1.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Experiment 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/300\n",
      "59/59 [==============================] - 24s 399ms/step - loss: 4.1487 - accuracy: 0.0979\n",
      "Epoch 2/300\n",
      "59/59 [==============================] - 24s 406ms/step - loss: 2.3010 - accuracy: 0.1125\n",
      "Epoch 3/300\n",
      "59/59 [==============================] - 22s 368ms/step - loss: 2.2977 - accuracy: 0.1189\n",
      "Epoch 4/300\n",
      "59/59 [==============================] - 21s 357ms/step - loss: 2.2947 - accuracy: 0.1208\n",
      "Epoch 5/300\n",
      "59/59 [==============================] - 21s 349ms/step - loss: 2.2924 - accuracy: 0.1235\n",
      "Epoch 6/300\n",
      "59/59 [==============================] - 20s 342ms/step - loss: 2.2873 - accuracy: 0.1281\n",
      "Epoch 7/300\n",
      "59/59 [==============================] - 20s 331ms/step - loss: 2.2828 - accuracy: 0.1323\n",
      "Epoch 8/300\n",
      "59/59 [==============================] - 19s 322ms/step - loss: 2.2783 - accuracy: 0.1359\n",
      "Epoch 9/300\n",
      "59/59 [==============================] - 19s 325ms/step - loss: 2.2723 - accuracy: 0.1411\n",
      "Epoch 10/300\n",
      "59/59 [==============================] - 20s 333ms/step - loss: 2.2642 - accuracy: 0.1430\n",
      "Epoch 11/300\n",
      "59/59 [==============================] - 20s 334ms/step - loss: 2.2558 - accuracy: 0.1501\n",
      "Epoch 12/300\n",
      "59/59 [==============================] - 20s 341ms/step - loss: 2.2489 - accuracy: 0.1521\n",
      "Epoch 13/300\n",
      "59/59 [==============================] - 20s 346ms/step - loss: 2.2333 - accuracy: 0.1676\n",
      "Epoch 14/300\n",
      "59/59 [==============================] - 20s 345ms/step - loss: 2.2226 - accuracy: 0.1716\n",
      "Epoch 15/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 2.2086 - accuracy: 0.1783\n",
      "Epoch 16/300\n",
      "59/59 [==============================] - 21s 350ms/step - loss: 2.1954 - accuracy: 0.1852\n",
      "Epoch 17/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 2.1783 - accuracy: 0.1932\n",
      "Epoch 18/300\n",
      "59/59 [==============================] - 21s 355ms/step - loss: 2.1581 - accuracy: 0.2059\n",
      "Epoch 19/300\n",
      "59/59 [==============================] - 21s 362ms/step - loss: 2.1408 - accuracy: 0.2156\n",
      "Epoch 20/300\n",
      "59/59 [==============================] - 21s 349ms/step - loss: 2.1130 - accuracy: 0.2270\n",
      "Epoch 21/300\n",
      "59/59 [==============================] - 20s 340ms/step - loss: 2.0915 - accuracy: 0.2370\n",
      "Epoch 22/300\n",
      "59/59 [==============================] - 20s 337ms/step - loss: 2.0700 - accuracy: 0.2465\n",
      "Epoch 23/300\n",
      "59/59 [==============================] - 20s 331ms/step - loss: 2.0423 - accuracy: 0.2543\n",
      "Epoch 24/300\n",
      "59/59 [==============================] - 19s 326ms/step - loss: 2.0156 - accuracy: 0.2680\n",
      "Epoch 25/300\n",
      "59/59 [==============================] - 19s 325ms/step - loss: 1.9823 - accuracy: 0.2809\n",
      "Epoch 26/300\n",
      "59/59 [==============================] - 20s 331ms/step - loss: 1.9598 - accuracy: 0.2921\n",
      "Epoch 27/300\n",
      "59/59 [==============================] - 19s 325ms/step - loss: 1.9296 - accuracy: 0.3057\n",
      "Epoch 28/300\n",
      "59/59 [==============================] - 19s 326ms/step - loss: 1.8924 - accuracy: 0.3156\n",
      "Epoch 29/300\n",
      "59/59 [==============================] - 20s 334ms/step - loss: 1.8659 - accuracy: 0.3310\n",
      "Epoch 30/300\n",
      "59/59 [==============================] - 21s 356ms/step - loss: 1.8170 - accuracy: 0.3481\n",
      "Epoch 31/300\n",
      "59/59 [==============================] - 22s 369ms/step - loss: 1.7924 - accuracy: 0.3592\n",
      "Epoch 32/300\n",
      "59/59 [==============================] - 21s 359ms/step - loss: 1.7535 - accuracy: 0.3698\n",
      "Epoch 33/300\n",
      "59/59 [==============================] - 25s 431ms/step - loss: 1.7249 - accuracy: 0.3841\n",
      "Epoch 34/300\n",
      "59/59 [==============================] - 23s 383ms/step - loss: 1.6876 - accuracy: 0.3981\n",
      "Epoch 35/300\n",
      "59/59 [==============================] - 22s 373ms/step - loss: 1.6481 - accuracy: 0.4128\n",
      "Epoch 36/300\n",
      "59/59 [==============================] - 23s 395ms/step - loss: 1.6136 - accuracy: 0.4288\n",
      "Epoch 37/300\n",
      "59/59 [==============================] - 22s 370ms/step - loss: 1.5697 - accuracy: 0.4428\n",
      "Epoch 38/300\n",
      "59/59 [==============================] - 25s 422ms/step - loss: 1.5447 - accuracy: 0.4535\n",
      "Epoch 39/300\n",
      "59/59 [==============================] - 27s 454ms/step - loss: 1.5177 - accuracy: 0.4661\n",
      "Epoch 40/300\n",
      "59/59 [==============================] - 23s 386ms/step - loss: 1.4732 - accuracy: 0.4781\n",
      "Epoch 41/300\n",
      "59/59 [==============================] - 22s 375ms/step - loss: 1.4449 - accuracy: 0.4873\n",
      "Epoch 42/300\n",
      "59/59 [==============================] - 23s 394ms/step - loss: 1.4081 - accuracy: 0.5016\n",
      "Epoch 43/300\n",
      "59/59 [==============================] - 23s 386ms/step - loss: 1.3704 - accuracy: 0.5189\n",
      "Epoch 44/300\n",
      "59/59 [==============================] - 22s 375ms/step - loss: 1.3376 - accuracy: 0.5278\n",
      "Epoch 45/300\n",
      "59/59 [==============================] - 24s 411ms/step - loss: 1.3154 - accuracy: 0.5343\n",
      "Epoch 46/300\n",
      "59/59 [==============================] - 24s 407ms/step - loss: 1.2833 - accuracy: 0.5470\n",
      "Epoch 47/300\n",
      "59/59 [==============================] - 22s 381ms/step - loss: 1.2533 - accuracy: 0.5579\n",
      "Epoch 48/300\n",
      "59/59 [==============================] - 22s 376ms/step - loss: 1.2247 - accuracy: 0.5670\n",
      "Epoch 49/300\n",
      "59/59 [==============================] - 22s 371ms/step - loss: 1.1879 - accuracy: 0.5791\n",
      "Epoch 50/300\n",
      "59/59 [==============================] - 21s 353ms/step - loss: 1.1518 - accuracy: 0.5944\n",
      "Epoch 51/300\n",
      "59/59 [==============================] - 22s 380ms/step - loss: 1.1322 - accuracy: 0.5982\n",
      "Epoch 52/300\n",
      "59/59 [==============================] - 24s 406ms/step - loss: 1.0983 - accuracy: 0.6161\n",
      "Epoch 53/300\n",
      "59/59 [==============================] - 22s 369ms/step - loss: 1.0731 - accuracy: 0.6243\n",
      "Epoch 54/300\n",
      "59/59 [==============================] - 22s 371ms/step - loss: 1.0554 - accuracy: 0.6276\n",
      "Epoch 55/300\n",
      "59/59 [==============================] - 26s 439ms/step - loss: 1.0348 - accuracy: 0.6349\n",
      "Epoch 56/300\n",
      "59/59 [==============================] - 22s 380ms/step - loss: 1.0094 - accuracy: 0.6423\n",
      "Epoch 57/300\n",
      "59/59 [==============================] - 22s 366ms/step - loss: 0.9791 - accuracy: 0.6560\n",
      "Epoch 58/300\n",
      "59/59 [==============================] - 25s 420ms/step - loss: 0.9715 - accuracy: 0.6565\n",
      "Epoch 59/300\n",
      "59/59 [==============================] - 25s 421ms/step - loss: 0.9565 - accuracy: 0.6659\n",
      "Epoch 60/300\n",
      "59/59 [==============================] - 24s 402ms/step - loss: 0.9334 - accuracy: 0.6683\n",
      "Epoch 61/300\n",
      "59/59 [==============================] - 25s 427ms/step - loss: 0.9055 - accuracy: 0.6816\n",
      "Epoch 62/300\n",
      "59/59 [==============================] - 23s 392ms/step - loss: 0.8793 - accuracy: 0.6878\n",
      "Epoch 63/300\n",
      "59/59 [==============================] - 21s 364ms/step - loss: 0.8775 - accuracy: 0.6905\n",
      "Epoch 64/300\n",
      "59/59 [==============================] - 21s 359ms/step - loss: 0.8410 - accuracy: 0.7022\n",
      "Epoch 65/300\n",
      "59/59 [==============================] - 22s 367ms/step - loss: 0.8308 - accuracy: 0.7065\n",
      "Epoch 66/300\n",
      "59/59 [==============================] - 24s 404ms/step - loss: 0.8130 - accuracy: 0.7129\n",
      "Epoch 67/300\n",
      "59/59 [==============================] - 25s 429ms/step - loss: 0.7941 - accuracy: 0.7200\n",
      "Epoch 68/300\n",
      "59/59 [==============================] - 22s 378ms/step - loss: 0.7745 - accuracy: 0.7274\n",
      "Epoch 69/300\n",
      "59/59 [==============================] - 24s 402ms/step - loss: 0.7548 - accuracy: 0.7361\n",
      "Epoch 70/300\n",
      "59/59 [==============================] - 23s 383ms/step - loss: 0.7468 - accuracy: 0.7377\n",
      "Epoch 71/300\n",
      "59/59 [==============================] - 24s 404ms/step - loss: 0.7392 - accuracy: 0.7387\n",
      "Epoch 72/300\n",
      "59/59 [==============================] - 24s 409ms/step - loss: 0.7123 - accuracy: 0.7511\n",
      "Epoch 73/300\n",
      "59/59 [==============================] - 22s 366ms/step - loss: 0.7188 - accuracy: 0.7466\n",
      "Epoch 74/300\n",
      "59/59 [==============================] - 25s 428ms/step - loss: 0.6994 - accuracy: 0.7557\n",
      "Epoch 75/300\n",
      "59/59 [==============================] - 24s 406ms/step - loss: 0.6805 - accuracy: 0.7614\n",
      "Epoch 76/300\n",
      "59/59 [==============================] - 22s 374ms/step - loss: 0.6697 - accuracy: 0.7632\n",
      "Epoch 77/300\n",
      "59/59 [==============================] - 23s 383ms/step - loss: 0.6612 - accuracy: 0.7677\n",
      "Epoch 78/300\n",
      "59/59 [==============================] - 23s 392ms/step - loss: 0.6451 - accuracy: 0.7751\n",
      "Epoch 79/300\n",
      "59/59 [==============================] - 22s 378ms/step - loss: 0.6286 - accuracy: 0.7798\n",
      "Epoch 80/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59/59 [==============================] - 23s 396ms/step - loss: 0.6193 - accuracy: 0.7829\n",
      "Epoch 81/300\n",
      "59/59 [==============================] - 25s 431ms/step - loss: 0.6200 - accuracy: 0.7840\n",
      "Epoch 82/300\n",
      "59/59 [==============================] - 26s 436ms/step - loss: 0.6096 - accuracy: 0.7853\n",
      "Epoch 83/300\n",
      "59/59 [==============================] - 24s 413ms/step - loss: 0.6102 - accuracy: 0.7867\n",
      "Epoch 84/300\n",
      "59/59 [==============================] - 25s 419ms/step - loss: 0.5856 - accuracy: 0.7941\n",
      "Epoch 85/300\n",
      "59/59 [==============================] - 22s 380ms/step - loss: 0.5848 - accuracy: 0.7945\n",
      "Epoch 86/300\n",
      "59/59 [==============================] - 23s 396ms/step - loss: 0.5677 - accuracy: 0.8001\n",
      "Epoch 87/300\n",
      "59/59 [==============================] - 21s 359ms/step - loss: 0.5603 - accuracy: 0.8001\n",
      "Epoch 88/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 0.5519 - accuracy: 0.8076\n",
      "Epoch 89/300\n",
      "59/59 [==============================] - 23s 391ms/step - loss: 0.5388 - accuracy: 0.8089\n",
      "Epoch 90/300\n",
      "59/59 [==============================] - 24s 401ms/step - loss: 0.5300 - accuracy: 0.8134\n",
      "Epoch 91/300\n",
      "59/59 [==============================] - 23s 386ms/step - loss: 0.5278 - accuracy: 0.8126\n",
      "Epoch 92/300\n",
      "59/59 [==============================] - 25s 422ms/step - loss: 0.5111 - accuracy: 0.8237\n",
      "Epoch 93/300\n",
      "59/59 [==============================] - 25s 425ms/step - loss: 0.5207 - accuracy: 0.8160\n",
      "Epoch 94/300\n",
      "59/59 [==============================] - 24s 406ms/step - loss: 0.5105 - accuracy: 0.8211\n",
      "Epoch 95/300\n",
      "59/59 [==============================] - 27s 451ms/step - loss: 0.4995 - accuracy: 0.8273\n",
      "Epoch 96/300\n",
      "59/59 [==============================] - 26s 444ms/step - loss: 0.4920 - accuracy: 0.8282\n",
      "Epoch 97/300\n",
      "59/59 [==============================] - 25s 422ms/step - loss: 0.4923 - accuracy: 0.8279\n",
      "Epoch 98/300\n",
      "59/59 [==============================] - 24s 402ms/step - loss: 0.4880 - accuracy: 0.8289\n",
      "Epoch 99/300\n",
      "59/59 [==============================] - 24s 406ms/step - loss: 0.4799 - accuracy: 0.8328\n",
      "Epoch 100/300\n",
      "59/59 [==============================] - 25s 418ms/step - loss: 0.4660 - accuracy: 0.8371\n",
      "Epoch 101/300\n",
      "59/59 [==============================] - 24s 413ms/step - loss: 0.4589 - accuracy: 0.8414\n",
      "Epoch 102/300\n",
      "59/59 [==============================] - 23s 396ms/step - loss: 0.4542 - accuracy: 0.8417\n",
      "Epoch 103/300\n",
      "59/59 [==============================] - 23s 394ms/step - loss: 0.4492 - accuracy: 0.8444\n",
      "Epoch 104/300\n",
      "59/59 [==============================] - 25s 432ms/step - loss: 0.4533 - accuracy: 0.8403\n",
      "Epoch 105/300\n",
      "59/59 [==============================] - 26s 434ms/step - loss: 0.4496 - accuracy: 0.8411\n",
      "Epoch 106/300\n",
      "59/59 [==============================] - 27s 451ms/step - loss: 0.4400 - accuracy: 0.8468\n",
      "Epoch 107/300\n",
      "59/59 [==============================] - 25s 428ms/step - loss: 0.4289 - accuracy: 0.8477\n",
      "Epoch 108/300\n",
      "59/59 [==============================] - 25s 421ms/step - loss: 0.4282 - accuracy: 0.8495\n",
      "Epoch 109/300\n",
      "59/59 [==============================] - 25s 428ms/step - loss: 0.4337 - accuracy: 0.8480\n",
      "Epoch 110/300\n",
      "59/59 [==============================] - 26s 444ms/step - loss: 0.4194 - accuracy: 0.8536\n",
      "Epoch 111/300\n",
      "59/59 [==============================] - 26s 437ms/step - loss: 0.4169 - accuracy: 0.8536\n",
      "Epoch 112/300\n",
      "59/59 [==============================] - 23s 391ms/step - loss: 0.4199 - accuracy: 0.8514\n",
      "Epoch 113/300\n",
      "59/59 [==============================] - 23s 397ms/step - loss: 0.4051 - accuracy: 0.8597\n",
      "Epoch 114/300\n",
      "59/59 [==============================] - 25s 423ms/step - loss: 0.4028 - accuracy: 0.8610\n",
      "Epoch 115/300\n",
      "59/59 [==============================] - 24s 408ms/step - loss: 0.3933 - accuracy: 0.8638\n",
      "Epoch 116/300\n",
      "59/59 [==============================] - 25s 416ms/step - loss: 0.4036 - accuracy: 0.8601\n",
      "Epoch 117/300\n",
      "59/59 [==============================] - 22s 377ms/step - loss: 0.4151 - accuracy: 0.8568\n",
      "Epoch 118/300\n",
      "59/59 [==============================] - 23s 394ms/step - loss: 0.3912 - accuracy: 0.8633\n",
      "Epoch 119/300\n",
      "59/59 [==============================] - 24s 414ms/step - loss: 0.3886 - accuracy: 0.8667\n",
      "Epoch 120/300\n",
      "59/59 [==============================] - 25s 420ms/step - loss: 0.3786 - accuracy: 0.8677\n",
      "Epoch 121/300\n",
      "59/59 [==============================] - 23s 381ms/step - loss: 0.3799 - accuracy: 0.8665\n",
      "Epoch 122/300\n",
      "59/59 [==============================] - 23s 390ms/step - loss: 0.3837 - accuracy: 0.8652\n",
      "Epoch 123/300\n",
      "59/59 [==============================] - 24s 398ms/step - loss: 0.3877 - accuracy: 0.8645\n",
      "Epoch 124/300\n",
      "59/59 [==============================] - 24s 412ms/step - loss: 0.3816 - accuracy: 0.8673\n",
      "Epoch 125/300\n",
      "59/59 [==============================] - 24s 404ms/step - loss: 0.3722 - accuracy: 0.8692\n",
      "Epoch 126/300\n",
      "59/59 [==============================] - 23s 390ms/step - loss: 0.3560 - accuracy: 0.8768\n",
      "Epoch 127/300\n",
      "59/59 [==============================] - 25s 421ms/step - loss: 0.3602 - accuracy: 0.8770\n",
      "Epoch 128/300\n",
      "59/59 [==============================] - 26s 439ms/step - loss: 0.3569 - accuracy: 0.8779\n",
      "Epoch 129/300\n",
      "59/59 [==============================] - 27s 460ms/step - loss: 0.3558 - accuracy: 0.8741\n",
      "Epoch 130/300\n",
      "59/59 [==============================] - 26s 435ms/step - loss: 0.3580 - accuracy: 0.8753\n",
      "Epoch 131/300\n",
      "59/59 [==============================] - 24s 400ms/step - loss: 0.3510 - accuracy: 0.8777\n",
      "Epoch 132/300\n",
      "59/59 [==============================] - 21s 354ms/step - loss: 0.3445 - accuracy: 0.8797\n",
      "Epoch 133/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 0.3537 - accuracy: 0.8780\n",
      "Epoch 134/300\n",
      "59/59 [==============================] - 23s 393ms/step - loss: 0.3391 - accuracy: 0.8824\n",
      "Epoch 135/300\n",
      "59/59 [==============================] - 26s 448ms/step - loss: 0.3464 - accuracy: 0.8792\n",
      "Epoch 136/300\n",
      "59/59 [==============================] - 26s 441ms/step - loss: 0.3416 - accuracy: 0.8812\n",
      "Epoch 137/300\n",
      "59/59 [==============================] - 25s 427ms/step - loss: 0.3507 - accuracy: 0.8782\n",
      "Epoch 138/300\n",
      "59/59 [==============================] - 28s 473ms/step - loss: 0.3338 - accuracy: 0.8866\n",
      "Epoch 139/300\n",
      "59/59 [==============================] - 30s 511ms/step - loss: 0.3389 - accuracy: 0.8827\n",
      "Epoch 140/300\n",
      "59/59 [==============================] - 26s 447ms/step - loss: 0.3235 - accuracy: 0.8859\n",
      "Epoch 141/300\n",
      "59/59 [==============================] - 23s 382ms/step - loss: 0.3278 - accuracy: 0.8850\n",
      "Epoch 142/300\n",
      "59/59 [==============================] - 21s 363ms/step - loss: 0.3367 - accuracy: 0.8821\n",
      "Epoch 143/300\n",
      "59/59 [==============================] - 23s 383ms/step - loss: 0.3283 - accuracy: 0.8865\n",
      "Epoch 144/300\n",
      "59/59 [==============================] - 24s 410ms/step - loss: 0.3246 - accuracy: 0.8861\n",
      "Epoch 145/300\n",
      "59/59 [==============================] - 23s 390ms/step - loss: 0.3182 - accuracy: 0.8913\n",
      "Epoch 146/300\n",
      "59/59 [==============================] - 22s 377ms/step - loss: 0.3184 - accuracy: 0.8881\n",
      "Epoch 147/300\n",
      "59/59 [==============================] - 25s 423ms/step - loss: 0.3121 - accuracy: 0.8902\n",
      "Epoch 148/300\n",
      "59/59 [==============================] - 23s 391ms/step - loss: 0.3118 - accuracy: 0.8927\n",
      "Epoch 149/300\n",
      "59/59 [==============================] - 26s 436ms/step - loss: 0.3238 - accuracy: 0.8866\n",
      "Epoch 150/300\n",
      "59/59 [==============================] - 25s 417ms/step - loss: 0.3081 - accuracy: 0.8921\n",
      "Epoch 151/300\n",
      "59/59 [==============================] - 23s 385ms/step - loss: 0.3146 - accuracy: 0.8907\n",
      "Epoch 152/300\n",
      "59/59 [==============================] - 24s 414ms/step - loss: 0.3065 - accuracy: 0.8915\n",
      "Epoch 153/300\n",
      "59/59 [==============================] - 22s 374ms/step - loss: 0.3079 - accuracy: 0.8929\n",
      "Epoch 154/300\n",
      "59/59 [==============================] - 21s 357ms/step - loss: 0.3038 - accuracy: 0.8912\n",
      "Epoch 155/300\n",
      "59/59 [==============================] - 23s 396ms/step - loss: 0.2944 - accuracy: 0.8986\n",
      "Epoch 156/300\n",
      "59/59 [==============================] - 29s 498ms/step - loss: 0.2912 - accuracy: 0.9001\n",
      "Epoch 157/300\n",
      "59/59 [==============================] - 25s 432ms/step - loss: 0.2889 - accuracy: 0.8981\n",
      "Epoch 158/300\n",
      "59/59 [==============================] - 34s 578ms/step - loss: 0.2907 - accuracy: 0.8988\n",
      "Epoch 159/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59/59 [==============================] - 35s 594ms/step - loss: 0.3013 - accuracy: 0.8954\n",
      "Epoch 160/300\n",
      "59/59 [==============================] - 37s 632ms/step - loss: 0.2768 - accuracy: 0.9019\n",
      "Epoch 161/300\n",
      "59/59 [==============================] - 32s 535ms/step - loss: 0.2986 - accuracy: 0.8957\n",
      "Epoch 162/300\n",
      "59/59 [==============================] - 40s 674ms/step - loss: 0.2958 - accuracy: 0.8965\n",
      "Epoch 163/300\n",
      "59/59 [==============================] - 36s 612ms/step - loss: 0.2926 - accuracy: 0.8987\n",
      "Epoch 164/300\n",
      "59/59 [==============================] - 37s 629ms/step - loss: 0.2908 - accuracy: 0.9001\n",
      "Epoch 165/300\n",
      "59/59 [==============================] - 35s 586ms/step - loss: 0.2864 - accuracy: 0.9012\n",
      "Epoch 166/300\n",
      "59/59 [==============================] - 40s 671ms/step - loss: 0.2889 - accuracy: 0.8997\n",
      "Epoch 167/300\n",
      "59/59 [==============================] - 29s 486ms/step - loss: 0.2941 - accuracy: 0.8979\n",
      "Epoch 168/300\n",
      "59/59 [==============================] - 26s 439ms/step - loss: 0.2892 - accuracy: 0.8991\n",
      "Epoch 169/300\n",
      "59/59 [==============================] - 24s 404ms/step - loss: 0.2892 - accuracy: 0.8989\n",
      "Epoch 170/300\n",
      "59/59 [==============================] - 20s 340ms/step - loss: 0.2722 - accuracy: 0.9051\n",
      "Epoch 171/300\n",
      "59/59 [==============================] - 20s 337ms/step - loss: 0.2689 - accuracy: 0.9085\n",
      "Epoch 172/300\n",
      "59/59 [==============================] - 20s 343ms/step - loss: 0.2762 - accuracy: 0.9030\n",
      "Epoch 173/300\n",
      "59/59 [==============================] - 21s 350ms/step - loss: 0.2767 - accuracy: 0.9035\n",
      "Epoch 174/300\n",
      "59/59 [==============================] - 21s 349ms/step - loss: 0.2750 - accuracy: 0.9030\n",
      "Epoch 175/300\n",
      "59/59 [==============================] - 21s 354ms/step - loss: 0.2772 - accuracy: 0.9046\n",
      "Epoch 176/300\n",
      "59/59 [==============================] - 21s 355ms/step - loss: 0.2720 - accuracy: 0.9046\n",
      "Epoch 177/300\n",
      "59/59 [==============================] - 21s 359ms/step - loss: 0.2781 - accuracy: 0.9014\n",
      "Epoch 178/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 0.2745 - accuracy: 0.9038\n",
      "Epoch 179/300\n",
      "59/59 [==============================] - 20s 345ms/step - loss: 0.2660 - accuracy: 0.9074\n",
      "Epoch 180/300\n",
      "59/59 [==============================] - 20s 336ms/step - loss: 0.2667 - accuracy: 0.9074\n",
      "Epoch 181/300\n",
      "59/59 [==============================] - 19s 330ms/step - loss: 0.2675 - accuracy: 0.9078\n",
      "Epoch 182/300\n",
      "59/59 [==============================] - 19s 325ms/step - loss: 0.2689 - accuracy: 0.9083\n",
      "Epoch 183/300\n",
      "59/59 [==============================] - 19s 323ms/step - loss: 0.2606 - accuracy: 0.9090\n",
      "Epoch 184/300\n",
      "59/59 [==============================] - 20s 331ms/step - loss: 0.2580 - accuracy: 0.9116\n",
      "Epoch 185/300\n",
      "59/59 [==============================] - 20s 336ms/step - loss: 0.2483 - accuracy: 0.9133\n",
      "Epoch 186/300\n",
      "59/59 [==============================] - 21s 349ms/step - loss: 0.2731 - accuracy: 0.9063\n",
      "Epoch 187/300\n",
      "59/59 [==============================] - 20s 347ms/step - loss: 0.2612 - accuracy: 0.9100\n",
      "Epoch 188/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 0.2593 - accuracy: 0.9096\n",
      "Epoch 189/300\n",
      "59/59 [==============================] - 21s 350ms/step - loss: 0.2574 - accuracy: 0.9123\n",
      "Epoch 190/300\n",
      "59/59 [==============================] - 21s 352ms/step - loss: 0.2441 - accuracy: 0.9157\n",
      "Epoch 191/300\n",
      "59/59 [==============================] - 21s 360ms/step - loss: 0.2455 - accuracy: 0.9157\n",
      "Epoch 192/300\n",
      "59/59 [==============================] - 21s 357ms/step - loss: 0.2564 - accuracy: 0.9116\n",
      "Epoch 193/300\n",
      "59/59 [==============================] - 21s 348ms/step - loss: 0.2617 - accuracy: 0.9090\n",
      "Epoch 194/300\n",
      "59/59 [==============================] - 20s 339ms/step - loss: 0.2604 - accuracy: 0.9119\n",
      "Epoch 195/300\n",
      "59/59 [==============================] - 20s 333ms/step - loss: 0.2622 - accuracy: 0.9085\n",
      "Epoch 196/300\n",
      "59/59 [==============================] - 19s 324ms/step - loss: 0.2615 - accuracy: 0.9083\n",
      "Epoch 197/300\n",
      "59/59 [==============================] - 19s 330ms/step - loss: 0.2607 - accuracy: 0.9097\n",
      "Epoch 198/300\n",
      "59/59 [==============================] - 20s 335ms/step - loss: 0.2532 - accuracy: 0.9126\n",
      "Epoch 199/300\n",
      "59/59 [==============================] - 20s 339ms/step - loss: 0.2457 - accuracy: 0.9149\n",
      "Epoch 200/300\n",
      "59/59 [==============================] - 20s 342ms/step - loss: 0.2508 - accuracy: 0.9125\n",
      "Epoch 201/300\n",
      "59/59 [==============================] - 20s 348ms/step - loss: 0.2412 - accuracy: 0.9148\n",
      "Epoch 202/300\n",
      "59/59 [==============================] - 21s 350ms/step - loss: 0.2451 - accuracy: 0.9129\n",
      "Epoch 203/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 0.2422 - accuracy: 0.9158\n",
      "Epoch 204/300\n",
      "59/59 [==============================] - 21s 356ms/step - loss: 0.2407 - accuracy: 0.9166\n",
      "Epoch 205/300\n",
      "59/59 [==============================] - 21s 354ms/step - loss: 0.2389 - accuracy: 0.9171\n",
      "Epoch 206/300\n",
      "59/59 [==============================] - 20s 344ms/step - loss: 0.2432 - accuracy: 0.9148\n",
      "Epoch 207/300\n",
      "59/59 [==============================] - 20s 338ms/step - loss: 0.2334 - accuracy: 0.9202\n",
      "Epoch 208/300\n",
      "59/59 [==============================] - 25s 425ms/step - loss: 0.2335 - accuracy: 0.9180\n",
      "Epoch 209/300\n",
      "59/59 [==============================] - 23s 392ms/step - loss: 0.2291 - accuracy: 0.9204\n",
      "Epoch 210/300\n",
      "59/59 [==============================] - 22s 365ms/step - loss: 0.2317 - accuracy: 0.9182\n",
      "Epoch 211/300\n",
      "59/59 [==============================] - 23s 394ms/step - loss: 0.2226 - accuracy: 0.9225\n",
      "Epoch 212/300\n",
      "59/59 [==============================] - 22s 375ms/step - loss: 0.2394 - accuracy: 0.9161\n",
      "Epoch 213/300\n",
      "59/59 [==============================] - 22s 377ms/step - loss: 0.2353 - accuracy: 0.9180\n",
      "Epoch 214/300\n",
      "59/59 [==============================] - 23s 397ms/step - loss: 0.2367 - accuracy: 0.9159\n",
      "Epoch 215/300\n",
      "59/59 [==============================] - 25s 414ms/step - loss: 0.2335 - accuracy: 0.9196\n",
      "Epoch 216/300\n",
      "59/59 [==============================] - 23s 391ms/step - loss: 0.2308 - accuracy: 0.9206\n",
      "Epoch 217/300\n",
      "59/59 [==============================] - 27s 465ms/step - loss: 0.2304 - accuracy: 0.9204\n",
      "Epoch 218/300\n",
      "59/59 [==============================] - 32s 536ms/step - loss: 0.2380 - accuracy: 0.9166\n",
      "Epoch 219/300\n",
      "59/59 [==============================] - 30s 510ms/step - loss: 0.2316 - accuracy: 0.9192\n",
      "Epoch 220/300\n",
      "59/59 [==============================] - 33s 565ms/step - loss: 0.2305 - accuracy: 0.9217\n",
      "Epoch 221/300\n",
      "59/59 [==============================] - 28s 482ms/step - loss: 0.2244 - accuracy: 0.9212\n",
      "Epoch 222/300\n",
      "59/59 [==============================] - 25s 414ms/step - loss: 0.2094 - accuracy: 0.9289\n",
      "Epoch 223/300\n",
      "59/59 [==============================] - 22s 378ms/step - loss: 0.2218 - accuracy: 0.9240\n",
      "Epoch 224/300\n",
      "59/59 [==============================] - 25s 416ms/step - loss: 0.2149 - accuracy: 0.9260\n",
      "Epoch 225/300\n",
      "59/59 [==============================] - 23s 397ms/step - loss: 0.2191 - accuracy: 0.9253\n",
      "Epoch 226/300\n",
      "59/59 [==============================] - 23s 398ms/step - loss: 0.2258 - accuracy: 0.9220\n",
      "Epoch 227/300\n",
      "59/59 [==============================] - 24s 408ms/step - loss: 0.2233 - accuracy: 0.9226\n",
      "Epoch 228/300\n",
      "59/59 [==============================] - 24s 409ms/step - loss: 0.2327 - accuracy: 0.9195\n",
      "Epoch 229/300\n",
      "59/59 [==============================] - 23s 395ms/step - loss: 0.2220 - accuracy: 0.9216\n",
      "Epoch 230/300\n",
      "59/59 [==============================] - 22s 380ms/step - loss: 0.2438 - accuracy: 0.9164\n",
      "Epoch 231/300\n",
      "59/59 [==============================] - 22s 370ms/step - loss: 0.2296 - accuracy: 0.9190\n",
      "Epoch 232/300\n",
      "59/59 [==============================] - 22s 370ms/step - loss: 0.2308 - accuracy: 0.9192\n",
      "Epoch 233/300\n",
      "59/59 [==============================] - 21s 357ms/step - loss: 0.2221 - accuracy: 0.9216\n",
      "Epoch 234/300\n",
      "59/59 [==============================] - 21s 358ms/step - loss: 0.2208 - accuracy: 0.9240\n",
      "Epoch 235/300\n",
      "59/59 [==============================] - 21s 360ms/step - loss: 0.2216 - accuracy: 0.9241\n",
      "Epoch 236/300\n",
      "59/59 [==============================] - 21s 359ms/step - loss: 0.2172 - accuracy: 0.9243\n",
      "Epoch 237/300\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "59/59 [==============================] - 21s 361ms/step - loss: 0.2040 - accuracy: 0.9272\n",
      "Epoch 238/300\n",
      "59/59 [==============================] - 22s 369ms/step - loss: 0.2155 - accuracy: 0.9254\n",
      "Epoch 239/300\n",
      "59/59 [==============================] - 22s 367ms/step - loss: 0.2053 - accuracy: 0.9286\n",
      "Epoch 240/300\n",
      "59/59 [==============================] - 22s 374ms/step - loss: 0.2154 - accuracy: 0.9258\n",
      "Epoch 241/300\n",
      "59/59 [==============================] - 23s 393ms/step - loss: 0.2114 - accuracy: 0.9268\n",
      "Epoch 242/300\n",
      "59/59 [==============================] - 24s 405ms/step - loss: 0.2062 - accuracy: 0.9288\n",
      "Epoch 243/300\n",
      "59/59 [==============================] - 23s 390ms/step - loss: 0.2044 - accuracy: 0.9278\n",
      "Epoch 244/300\n",
      "59/59 [==============================] - 23s 381ms/step - loss: 0.2210 - accuracy: 0.9228\n",
      "Epoch 245/300\n",
      "59/59 [==============================] - 22s 368ms/step - loss: 0.2121 - accuracy: 0.9268\n",
      "Epoch 246/300\n",
      "59/59 [==============================] - 22s 364ms/step - loss: 0.2118 - accuracy: 0.9270\n",
      "Epoch 247/300\n",
      "59/59 [==============================] - 21s 360ms/step - loss: 0.2147 - accuracy: 0.9269\n",
      "Epoch 248/300\n",
      "59/59 [==============================] - 21s 364ms/step - loss: 0.2051 - accuracy: 0.9289\n",
      "Epoch 249/300\n",
      "59/59 [==============================] - 22s 365ms/step - loss: 0.2146 - accuracy: 0.9263\n",
      "Epoch 250/300\n",
      "59/59 [==============================] - 22s 371ms/step - loss: 0.2160 - accuracy: 0.9258\n",
      "Epoch 251/300\n",
      "59/59 [==============================] - 23s 382ms/step - loss: 0.2174 - accuracy: 0.9224\n",
      "Epoch 252/300\n",
      "59/59 [==============================] - 23s 396ms/step - loss: 0.2038 - accuracy: 0.9287\n",
      "Epoch 253/300\n",
      "59/59 [==============================] - 23s 389ms/step - loss: 0.2096 - accuracy: 0.9283\n",
      "Epoch 254/300\n",
      "59/59 [==============================] - 22s 381ms/step - loss: 0.2007 - accuracy: 0.9297\n",
      "Epoch 255/300\n",
      "59/59 [==============================] - 22s 373ms/step - loss: 0.2011 - accuracy: 0.9310\n",
      "Epoch 256/300\n",
      "59/59 [==============================] - 22s 366ms/step - loss: 0.2025 - accuracy: 0.9308\n",
      "Epoch 257/300\n",
      "59/59 [==============================] - 21s 361ms/step - loss: 0.2010 - accuracy: 0.9295\n",
      "Epoch 258/300\n",
      "59/59 [==============================] - 21s 362ms/step - loss: 0.1923 - accuracy: 0.9324\n",
      "Epoch 259/300\n",
      "59/59 [==============================] - 22s 373ms/step - loss: 0.2117 - accuracy: 0.9254\n",
      "Epoch 260/300\n",
      "59/59 [==============================] - 22s 377ms/step - loss: 0.2010 - accuracy: 0.9314\n",
      "Epoch 261/300\n",
      "59/59 [==============================] - 21s 348ms/step - loss: 0.2073 - accuracy: 0.9274\n",
      "Epoch 262/300\n",
      "59/59 [==============================] - 21s 357ms/step - loss: 0.2096 - accuracy: 0.9276\n",
      "Epoch 263/300\n",
      "59/59 [==============================] - 20s 346ms/step - loss: 0.2023 - accuracy: 0.9304\n",
      "Epoch 264/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 0.2020 - accuracy: 0.9301\n",
      "Epoch 265/300\n",
      "59/59 [==============================] - 21s 355ms/step - loss: 0.1880 - accuracy: 0.9359\n",
      "Epoch 266/300\n",
      "59/59 [==============================] - 21s 363ms/step - loss: 0.1912 - accuracy: 0.9337\n",
      "Epoch 267/300\n",
      "59/59 [==============================] - 21s 358ms/step - loss: 0.1976 - accuracy: 0.9312\n",
      "Epoch 268/300\n",
      "59/59 [==============================] - 21s 352ms/step - loss: 0.2075 - accuracy: 0.9294\n",
      "Epoch 269/300\n",
      "59/59 [==============================] - 20s 343ms/step - loss: 0.1971 - accuracy: 0.9322\n",
      "Epoch 270/300\n",
      "59/59 [==============================] - 20s 335ms/step - loss: 0.2069 - accuracy: 0.9283\n",
      "Epoch 271/300\n",
      "59/59 [==============================] - 20s 334ms/step - loss: 0.1987 - accuracy: 0.9306\n",
      "Epoch 272/300\n",
      "59/59 [==============================] - 19s 330ms/step - loss: 0.1927 - accuracy: 0.9345\n",
      "Epoch 273/300\n",
      "59/59 [==============================] - 20s 338ms/step - loss: 0.2107 - accuracy: 0.9287\n",
      "Epoch 274/300\n",
      "59/59 [==============================] - 20s 342ms/step - loss: 0.2006 - accuracy: 0.9305\n",
      "Epoch 275/300\n",
      "59/59 [==============================] - 20s 344ms/step - loss: 0.1996 - accuracy: 0.9299\n",
      "Epoch 276/300\n",
      "59/59 [==============================] - 20s 343ms/step - loss: 0.2037 - accuracy: 0.9295\n",
      "Epoch 277/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 0.1965 - accuracy: 0.9323\n",
      "Epoch 278/300\n",
      "59/59 [==============================] - 21s 351ms/step - loss: 0.1945 - accuracy: 0.9332\n",
      "Epoch 279/300\n",
      "59/59 [==============================] - 21s 353ms/step - loss: 0.1988 - accuracy: 0.9314\n",
      "Epoch 280/300\n",
      "59/59 [==============================] - 21s 361ms/step - loss: 0.1949 - accuracy: 0.9314\n",
      "Epoch 281/300\n",
      "59/59 [==============================] - 22s 366ms/step - loss: 0.1890 - accuracy: 0.9338\n",
      "Epoch 282/300\n",
      "59/59 [==============================] - 21s 361ms/step - loss: 0.2010 - accuracy: 0.9305\n",
      "Epoch 283/300\n",
      "59/59 [==============================] - 21s 362ms/step - loss: 0.1943 - accuracy: 0.9328\n",
      "Epoch 284/300\n",
      "59/59 [==============================] - 21s 356ms/step - loss: 0.1880 - accuracy: 0.9348\n",
      "Epoch 285/300\n",
      "59/59 [==============================] - 22s 369ms/step - loss: 0.1952 - accuracy: 0.9319\n",
      "Epoch 286/300\n",
      "59/59 [==============================] - 21s 360ms/step - loss: 0.1936 - accuracy: 0.9338\n",
      "Epoch 287/300\n",
      "59/59 [==============================] - 22s 368ms/step - loss: 0.1910 - accuracy: 0.9339\n",
      "Epoch 288/300\n",
      "59/59 [==============================] - 22s 365ms/step - loss: 0.1800 - accuracy: 0.9374\n",
      "Epoch 289/300\n",
      "59/59 [==============================] - 23s 381ms/step - loss: 0.1864 - accuracy: 0.9366\n",
      "Epoch 290/300\n",
      "59/59 [==============================] - 22s 374ms/step - loss: 0.1835 - accuracy: 0.9379\n",
      "Epoch 291/300\n",
      "59/59 [==============================] - 22s 366ms/step - loss: 0.1826 - accuracy: 0.9373\n",
      "Epoch 292/300\n",
      "59/59 [==============================] - 21s 353ms/step - loss: 0.1961 - accuracy: 0.9316\n",
      "Epoch 293/300\n",
      "59/59 [==============================] - 21s 357ms/step - loss: 0.1939 - accuracy: 0.9319\n",
      "Epoch 294/300\n",
      "59/59 [==============================] - 21s 361ms/step - loss: 0.1925 - accuracy: 0.9324\n",
      "Epoch 295/300\n",
      "59/59 [==============================] - 22s 366ms/step - loss: 0.1903 - accuracy: 0.9351\n",
      "Epoch 296/300\n",
      "59/59 [==============================] - 21s 364ms/step - loss: 0.1915 - accuracy: 0.9344\n",
      "Epoch 297/300\n",
      "59/59 [==============================] - 20s 347ms/step - loss: 0.1793 - accuracy: 0.9377\n",
      "Epoch 298/300\n",
      "59/59 [==============================] - 20s 340ms/step - loss: 0.1750 - accuracy: 0.9385\n",
      "Epoch 299/300\n",
      "59/59 [==============================] - 20s 337ms/step - loss: 0.1855 - accuracy: 0.9343\n",
      "Epoch 300/300\n",
      "59/59 [==============================] - 19s 329ms/step - loss: 0.1869 - accuracy: 0.9337\n",
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "batch_normalization (BatchNo (None, 26, 26, 14)        56        \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 13, 13, 14)        0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 2366)              0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1024)              2423808   \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 10)                10250     \n",
      "=================================================================\n",
      "Total params: 2,434,114\n",
      "Trainable params: 2,434,086\n",
      "Non-trainable params: 28\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.layers import BatchNormalization\n",
    "from tensorflow.keras.layers import Dense\n",
    "from tensorflow.keras.layers import Dropout\n",
    "\n",
    "model_2 = Sequential()\n",
    "\n",
    "model_2.add(BatchNormalization())\n",
    "model_2.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model_2.add(Flatten())\n",
    "model_2.add(Dense(1024, activation='relu', kernel_initializer=initializer))\n",
    "model_2.add(Dropout(0.2))\n",
    "model_2.add(Dense(10, activation='softmax', kernel_initializer=initializer))\n",
    "\n",
    "model_2.compile(optimizer='adam', \n",
    "               loss='sparse_categorical_crossentropy', \n",
    "               metrics = ['accuracy'])\n",
    "\n",
    "history_data2 = model_2.fit(x=train_data, \n",
    "                                     y=train_labels, \n",
    "                                     epochs=300, \n",
    "                                     batch_size=1024)\n",
    "\n",
    "model_2.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3. GRAPH"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.1. Model 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x19cff23a130>"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(history_data1.history['loss'], label = \"train_loss\")\n",
    "plt.xlabel('iteration')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x19cff4201f0>"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEGCAYAAABy53LJAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/Il7ecAAAACXBIWXMAAAsTAAALEwEAmpwYAAAsbElEQVR4nO3deXxV1d3v8c+PJBDmIQmIgIKKIFoRDYNKHR8VrIoDWmy1znRQq+3tbenTViv23upjB7XXolh5rD4oDi2KLUJxoM5CQEQQkAgIQZBAmCEkJ+d3/9g7eDgkcE7IIYn7+3698uLstYfzWxw4v6y19l7L3B0REZFUNWvoAEREpGlR4hARkbQocYiISFqUOEREJC1KHCIikpbshg7gYMjPz/eePXs2dBgiIk3KnDlz1rt7QXJ5JBJHz549KSoqaugwRESaFDP7rKZydVWJiEhalDhERCQtShwiIpKWSIxxiMjBU1lZSUlJCeXl5Q0diqQoNzeX7t27k5OTk9LxGU0cZjYMeADIAv7i7vck7T8NuB84Hhjl7s8n7JsGDAHecvcLarj2g8D17t4mczUQkXSVlJTQtm1bevbsiZk1dDiyH+7Ohg0bKCkpoVevXimdk7GuKjPLAh4ChgP9gCvNrF/SYSuBa4GnarjEfcDVtVy7EOhYb8GKSL0pLy8nLy9PSaOJMDPy8vLSaiFmcoxjEFDs7svcvQKYBIxIPMDdV7j7fCCefLK7vwpsTS4PE9J9wE8zErWIHDAljaYl3c8rk4mjG7AqYbskLDtQtwBT3H3Nvg4ys9FmVmRmRaWlpfXwttJQpi1Yy4LVmxs6DBEJNam7qszsUOBy4E/7O9bdx7t7obsXFhTs9eCjNBFfbCnne/8zhwv+9BYvffh5Q4cjImQ2cawGeiRsdw/LDsQA4Cig2MxWAK3MrPgArymN2L+XfNlafHGeEofs36ZNm/jzn/+c9nnnn38+mzZtqv+AvoIymThmA73NrJeZNQdGAVMO5ILu/k93P8Tde7p7T2CHux9VD7FKI/DW0vXsrKjC3Xno9WIe+fenTHh7OYe0y+XSE7sxd+VGtGKl7E9tiSMWi+3zvKlTp9KhQ4cMRXXg9hf/wZSx23HdPWZmtwDTCW7HneDuC81sLFDk7lPMbCAwmeAOqQvN7C53PxbAzN4E+gJtzKwEuMHdp2cqXjl4dsWqmDRrFSNOOJQOrZoDsGTtVq567H1uPeso+hzSlvumL9l9/DcLe3B8j/b8fe5qPtuwg575rRsqdEnTXS8t5OPPt9TrNfsd2o47Lzy21v1jxozh008/5YQTTiAnJ4fc3Fw6duzI4sWL+eSTT7j44otZtWoV5eXl3HbbbYwePRr4ck67bdu2MXz4cIYOHco777xDt27dePHFF2nZsmWN7/foo48yfvx4KioqOOqoo3jyySdp1aoVX3zxBd/73vdYtmwZAOPGjeOUU07hiSee4He/+x1mxvHHH8+TTz7JtddeywUXXMDIkSMBaNOmDdu2bWPmzJn86le/Sin+adOm8Z//+Z9UVVWRn5/PjBkz6NOnD++88w4FBQXE43GOPvpo3n33XQ60+z6jz3G4+1RgalLZHQmvZxN0YdV07tdTuL6e4Wjk1m0t57dTF3PD0F48NWslt53dm8kfrOaelxfzwcqN/PbS43luziqWr98OwPNzSsjJakbfQ9ry39cNZMbHX3D2MV3YsrMSgLkrN9IzvzX3v/IJxx7annP6dWnI6kkjdM8997BgwQLmzZvHzJkz+cY3vsGCBQt2P6MwYcIEOnXqxM6dOxk4cCCXXXYZeXl5e1xj6dKlPP300zz66KNcccUV/O1vf+Oqq66q8f0uvfRSbrrpJgB++ctf8thjj3Hrrbfywx/+kNNPP53JkydTVVXFtm3bWLhwIb/5zW945513yM/Pp6ysbL/1mTt37n7jj8fj3HTTTbzxxhv06tWLsrIymjVrxlVXXcXEiRO5/fbbeeWVV+jfv/8BJw3Qk+NyANZtKeeeaYspXreNv1xTSHlFnIff+JSfD+9L29zgCdTJc1cz+YPV/POjNVTE4qzdXM7sFWW0bZHNC/M+Z9byMj7fHNw/3iK7GWs2l5OTZTx5w2C6tm/Jd07uCcAh7XJp2yKbWcvLOOXIfO5/ZSmH57Xi7L6dadbMeG/ZBlpkN2PAYXq8pzHZV8vgYBk0aNAeD7Y9+OCDTJ48GYBVq1axdOnSvRJHr169OOGEEwA46aSTWLFiRa3XX7BgAb/85S/ZtGkT27Zt47zzzgPgtdde44knngAgKyuL9u3b88QTT3D55ZeTn58PQKdOneol/tLSUk477bTdx1Vf9/rrr2fEiBHcfvvtTJgwgeuuu26/75cKJQ5Ji7tjZlTFnZufmstHqzcTd/jp8/M5qqANT72/kvKKKsZefBxtWmQz4+MvaGZQEYtzWKdWvLZ4HR1a5TDxxsG8OO9zFq3Zwul9Cnh61iq+e/qRrCrbwYX9uzLkiD3/I2c1M4b2zmfmklL6HtIWgM827OCt4vWccFgHbnh8NuWxOPdedjwjT6qxESsR1br1l12bM2fO5JVXXuHdd9+lVatWnHHGGTU++NaiRYvdr7Oysti5c2et17/22mt54YUX6N+/P48//jgzZ85MO8bs7Gzi8eBxtng8TkVFxQHFX61Hjx506dKF1157jVmzZjFx4sS0Y6sx3nq5ijRp5ZVVVMWd1i2+/OfwdvF6Xl20jjP6FDDkiDy2lFfywgeruXfaYnp0bMXWXTFKt+7iD1f0Z/POSu566WPe+XQDLXOy+PsHq5nx8Rf85ZpC5qzcyM1nHMVJPTsyoEcH/vLmci4v7M7hea059tD2QJCMzuzTmVOPyt8jhmRn9unMywvWcu+0JRxR0JotO2PcNukDTjq8I9srqujXtR2/emEBQ47oxJadMV78cDX9urZjxAndeOnDz9lSXsm3Bx/OjooYOyqqyG8TfDm8OG81T7z7GQN6dODKwYdxZIF6QJuytm3bsnXrXs8OA7B582Y6duxIq1atWLx4Me+9994Bv9/WrVvp2rUrlZWVTJw4kW7dgsfVzj77bMaNG8ftt9++u6vqrLPO4pJLLuHHP/4xeXl5lJWV0alTJ3r27MmcOXO44oormDJlCpWVlWnFP2TIEH7wgx+wfPny3V1V1a2OG2+8kauuuoqrr76arKysA64vKHF8Jbg7n28up03zbOav3sTRXdrSpV0uzxatonXzbMq272LTjkoq486Jh3WgIhbn4zVb+Fq39gzq1YlbnvqAxWu3cN/I/qzdUs6uWJx7pi5ie0UVE95eTo9OLSnbVkGLnCx65rXm6C5tyc3J4tSj8rhkQDeq4s7E91dSvG4b/zXyeDq3bcEPJs7lxr8W4Q4X9O9K30PaAfCT8/rsFb+Zce6xh+y3nqf3Cfpmd1ZWcd2pvRjcqxN3vLiAVxat4+Qj8vjdFf35j9//m9smzWPpF1vZUh4jJ8v49yel/H1ucCf4m5+sZ87KjWzfFeNr3dqzcUcFpVt3kdWsGR+VbObJ9z7jgVED6NAqZ69WT7Ut5ZW0y01tMjg5+PLy8jj11FM57rjjaNmyJV26fDkONmzYMB5++GGOOeYY+vTpw5AhQw74/e6++24GDx5MQUEBgwcP3p20HnjgAUaPHs1jjz1GVlYW48aN4+STT+YXv/gFp59+OllZWQwYMIDHH3+cm266iREjRtC/f3+GDRu2RysjUW3xFxQUMH78eC699FLi8TidO3dmxowZAFx00UVcd9119dZNBWBRuL2xsLDQm/IKgMtKt/HKoi848bCOfLByEycfmcftz8wjy4xvHN+Vv88tYcWGHbuPP6ZrO741+DB+9cKCva5lBokfeZd2Lfhiy669jju0fS7PfPdk7pyykNkryjBgS3mM/7lhMEN75+91/NvF67l32mKevGEw7Vvm8NuXF/HIv5dxw9Be/OqC5CnK6u7PM4vp3rEVF/U/dHdZ8bpt5LVuTsfWzXlx3mpuf2YeLXOyeOL6QVz/+Gy2lMe4/tReVMXj/GP+Go4saMOuWBXL1m/HHbbtivGPW4eS36YFl417h9Wbgm6JX13QDwO+c/LhZGcFd64/+e4Kfv3Sx4z79okpJbsoWrRoEcccc0xDhyGhoqIifvSjH/Hmm2/u87iaPjczm+PuhcnHKnEcZO7Ojoqq3V0y5ZVVzFxSyooN2xl6VD7tW+Yw57ONvLG0lMG9OvGP+Wt4c+n6Pa6R1SyYV6Z/9/bMXbmJru1z+f4ZR7K1PEasyvnjK58AcPIReVxyYjcK2rRgYK9OVMTi/PT5D2mRk8X/ufg43l9exq1PfUB2lvGnKwewZnM5X++dT1YzI79NC3JzsnB3dsXiu2MaM6xvSvPabNsV46UPP+eSAd3Izamf5nGq/v1JKS2ymzHkiDxmLS9j887Kve6+qoo7FbE4KzZsZ/n67Zz/ta4AfFq6jdcXr+PZolV88sU2AC4Z0I0/XNGfjTsqOeO+19m6K0arnCxm/u8zKWjbYq/3jzoljsbjnnvuYdy4cUycOJGhQ4fu81gljiQHM3FsKa/k03XbiDu8v3wD7rByww4Wr93CiYd35MV5n1O2vYKhR+XTo1NL/jl/DVvK936wp0V2M3bF4uS1bs71Q3tx3rFdmLZgLR1bN+euKR9z7ak9+fnwvry3rIyju7Qhr82XX2DPFq0iu1nQGmmRve8v7Xc/3UB5ZRVn9u1c738XTdny9duZvnAt23fF+NNrxdx10bEsXbeVp2et4g9X9Oe2SfO497Kv8c2BhzV0qI3OVzVx3Hzzzbz99tt7lN1222312gXUkJQ4kmQ6cbg70xasZdLsVcxaXsbOyqo99ufmNOOQdrms2LCDs/t2pneXtsz4eC0lG3dyTr8ufGvQYRzZuQ1vLl1PrCrO0Ye0pV/Xdry+eB2nHJlP+1Z79qdv3F5Bh1Y5moH0IHB3rnt8Nm8tXU/cne+c3JM7L+zHKfe8Rv/uHXj46pMaOsRGZ9GiRfTtm1rLVBoHd2fx4sUpJw4NjtfB2s3lzF25kYWfb6ZoxUZ2VFTx0erNHJ7XistO6sbJR+RTUVXFmX0606p5NlnNDHdnZdkOjgjv2BkzvO/uW1urJd9GOjzsPknWsXXzzFVO9mBm3P/NE3jg1aUsXrOV2/+jN2bGGX0KeOnDNVRWxcnJalJzhWZcbm4uGzZs0JocTUT1Qk65ubkpn6PEUQd/nlnME+9+BsCxh7Yjq5nxm4uPY9TAHrsHUfdmu5PG7hL9p2oSOrRqvteDbGf26czTs1bx3rINfL23Zl9O1L17d0pKStByBk1H9dKxqVLiqIMdFUFX1L9+dBpHd2nbwNFIQzjt6AJaNc/i5QVrlTiS5OTkpLwEqTRNamPXQawqeApaSSO6cnOyOLNvZ6YvWEvRirJ6m7V3ydqtjJv5KRf+6S3+++3l9XJNkfqmFkcdVMad7Cx1M0XdRf0P5Z/z1zDy4Xe5ZEA3dlZUUVy6jc5tW/D5pp1UVgX/Ts7s05krBx1G9441z65a7aPVm7n6sfeprHLatsjmD//6hEtP7E77lnrYUBoXJY46iFXFyWmmxlrUnduvCy/f9nWenrWSJ979jEPa5XJct/as21pOv0Pb0TInm807K5n4/mc8/s6KlK7ZM68VT48eQtn2Cr7x4FsMu/8NWjWv/+dgWjXPpld+a7Kb6Regr7qfDe9Ll3apD3ynQomjDmJVanFIcHPDMV3bcddFx/LtwYdzVOc2ux/OTLR+2y6mLVjLjop9L8TTzIzzv9aVru1b0rV9S34+vC/zM7TW+uYdlcxbtQnnq387ftSVJz0eUB+UOOog6KpSi0MCZkafQ2of78pv04Krhhye9nW/e/qRBxKWSMZk9NvPzIaZ2RIzKzazMTXsP83M5ppZzMxGJu2bZmabzOwfSeUTw2suMLMJZnbQO4Cr4nE18UUksjKWOMwsC3gIGA70A640s+TZ7lYC1wJP1XCJ+4CrayifSLCk7NeAlsCN9RRyyiqrXIlDRCIrky2OQUCxuy9z9wpgEjAi8QB3X+Hu84F48snu/iqw16T67j7VQ8Asall6NpNielpYRCIsk99+3YBVCdslYVm9CLuorgam1bJ/tJkVmVlRfT/BGtPtuCISYU351+Y/A2+4e42TzLv7eHcvdPfC+licPVHQVdWU/+pEROouk3dVrQZ6JGx3D8sOmJndCRQA362P66VLg+MiEmWZ/LV5NtDbzHqZWXNgFDDlQC9qZjcC5wFXuvteYyMHg57jEJEoy1jicPcYcAswHVgEPOvuC81srJldBGBmA82sBLgceMTMFlafb2ZvAs8BZ5tZiZmdF+56GOgCvGtm88zsjkzVoTaVcQ2Oi0h0ZfQBQHefCkxNKrsj4fVsarkryt2/Xkt5gz+0GNPtuCISYfq1uQ6Cyev0Vyci0aRvvzrQ4LiIRJkSRx1ocFxEokyJow40OC4iUaZvvzrQ4LiIRJkSR5rcPZxyRH91IhJN+vZLU1U8WPhGLQ4RiSoljjTFqhOHBsdFJKKUONJUWRXMcqI1x0UkqvTtl6ZYlVocIhJtShxpqowHLQ4NjotIVOnbL00aHBeRqFPiSNPuriolDhGJKCWONO0eHFdXlYhElL790qTbcUUk6jKaOMxsmJktMbNiMxtTw/7TzGyumcXMbGTSvmlmtsnM/pFU3svM3g+v+Uy4uuBBU93i0JrjIhJVGfv2M7Ms4CFgONAPuNLM+iUdthK4FniqhkvcB1xdQ/m9wB/d/ShgI3BDfcWcCg2Oi0jUZfLX5kFAsbsvc/cKYBIwIvEAd1/h7vOBvdYOd/dXga2JZWZmwFnA82HRX4GL6z/02lXqOQ4RibhMJo5uwKqE7ZKw7EDkAZvC9czr65ppiWlwXEQi7iv77Wdmo82syMyKSktL6+26MXVViUjEZTJxrAZ6JGx3D8sOxAagg5ll7++a7j7e3QvdvbCgoOAA3/ZLuwfH1eIQkYjK5LffbKB3eBdUc2AUMOVALujuDrwOVN+BdQ3w4gFFmSYNjotI1GUscYTjELcA04FFwLPuvtDMxprZRQBmNtDMSoDLgUfMbGH1+Wb2JvAccLaZlZjZeeGunwE/NrNigjGPxzJVh5pocFxEoi57/4fUnbtPBaYmld2R8Ho2QXdTTed+vZbyZQR3bDWIWFyD4yISbfr2S5PmqhKRqFPiSJPmqhKRqNO3X5qqB8ez1OIQkYhS4khTpSY5FJGIU+JIU0xrjotIxOnbL01ac1xEok6JI02Vuh1XRCJO335pqtLtuCIScUocaarUXVUiEnFKHGmKVcXJbmYES4OIiESPEkeaYnHXwLiIRJoSR5oqq+K6FVdEIk3fgGmqUotDRCJOiSNNlVVOllocIhJh+gZMU6wqTo5aHCISYUocaYo7NNMdVSISYRlNHGY2zMyWmFmxmY2pYf9pZjbXzGJmNjJp3zVmtjT8uSah/Eoz+8jM5pvZNDPLz2Qdkrk76qkSkSjL2FegmWUBDwHDgX7AlWbWL+mwlcC1wFNJ53YC7gQGE6z2d6eZdTSzbOAB4Ex3Px6YT7A87UETd1eLQ0QiLZO/Ow8Cit19mbtXAJOAEYkHuPsKd58PxJPOPQ+Y4e5l7r4RmAEMAyz8aW3BE3jtgM8zWIe9qKtKRKIuk4mjG7AqYbskLKvzue5eCXwf+IggYfQDHqvpAmY22syKzKyotLQ03dhrFXdHeUNEoqxJ9dabWQ5B4hgAHErQVfXzmo519/HuXujuhQUFBfUWg6vFISIRl8nEsRrokbDdPSw7kHNPAHD3T93dgWeBUw440jQEYxwH8x1FRBqXTCaO2UBvM+tlZs2BUcCUFM+dDpwbDoh3BM4Ny1YD/cysuglxDrConuPeJw2Oi0jUZWfqwu4eM7NbCL7ws4AJ7r7QzMYCRe4+xcwGApOBjsCFZnaXux/r7mVmdjdB8gEY6+5lAGZ2F/CGmVUCnxHclXXQxB3NjCsikZaxxAHg7lOBqUlldyS8nk3QDVXTuROACTWUPww8XL+Rps7VVSUiEbffriozu9DMmtQgeibpdlwRibpUEsI3gaVm9l9m1jfTATV2GhwXkajbb+Jw96sIbn/9FHjczN4Nn5Fom/HoGiGNcYhI1KXUBeXuW4DnCZ7+7gpcAsw1s1szGFujpDEOEYm6VMY4LjKzycBMIAcY5O7Dgf7A/8pseI2PbscVkahL5a6qy4A/uvsbiYXuvsPMbshMWI1XPK7BcRGJtlQSx6+BNdUbZtYS6BJOUPhqpgJrrDRXlYhEXSpjHM+x5+y1VWFZJGmuKhGJulQSR3Y4LToA4evmmQupcavSQk4iEnGpfAWWmtlF1RtmNgJYn7mQGjcNjotI1KUyxvE9YKKZ/T+CRZRWAd/JaFSNmJ7jEJGo22/icPdPgSFm1ibc3pbxqBoxdydLeUNEIiylSQ7N7BvAsUBu9W/b7j42g3E1WuqqEpGoS+UBwIcJ5qu6laCr6nLg8AzH1WjF4+qqEpFoS2Vw/BR3/w6w0d3vAk4Gjs5sWI2XJjkUkahLJXGUh3/uMLNDgUqC+aoiSc9xiEjUpZI4XjKzDsB9wFxgBfBUKhc3s2FmtsTMis1sTA37TzOzuWYWM7ORSfuuMbOl4c81CeXNzWy8mX1iZovN7LJUYqkvcT3HISIRt8/B8XABp1fdfRPwNzP7B5Dr7pv3d2EzywIeIlgXvASYbWZT3P3jhMNWEiz9+pOkczsBdwKFgANzwnM3Ar8A1rn70WF8nVKqaT0JphxRi0NEomufvzu7e5zgy796e1cqSSM0CCh292Xh0+aTgBFJ11/h7vPZc0oTgPOAGe5eFiaLGcCwcN/1wG+r43P3g/oworqqRCTqUul0edXMLrP0f83uRvCwYLWSsKzO54ZdZgB3h11cz5lZl5ouEC42VWRmRaWlpWmGXjsNjotI1KWSOL5LMKnhLjPbYmZbzWxLhuOqTTbQHXjH3U8E3gV+V9OB7j7e3QvdvbCgoKDeAtCa4yISdaksHdvW3Zu5e3N3bxdut0vh2quBHgnb3cOyVNR27gZgB/D3sPw54MQUr1kvNK26iETdfp8cN7PTaipPXtipBrOB3mbWi+BLfxTwrRTjmg78XzPrGG6fC/zc3d3MXgLOAF4DzgY+rvkSmaExDhGJulSmHPnfCa9zCQa95wBn7eskd4+Z2S0ESSALmODuC81sLFDk7lPMbCAwGegIXGhmd7n7se5eZmZ3EyQfgLHuXha+/hnwpJndD5QC16VU03qiMQ4RibpUJjm8MHHbzHoA96dycXefCkxNKrsj4fVsgm6oms6dAEyoofwzoMZW0MGguapEJOrq8ihbCXBMfQfSVGhadRGJulTGOP5E8BAeBInmBIInyCPJ1VUlIhGXyhhHUcLrGPC0u7+doXgaPd2OKyJRl0rieB4od/cqCKYSMbNW7r4js6E1ThocF5GoS+nJcaBlwnZL4JXMhNP4xeOaq0pEoi2VxJGbuFxs+LpV5kJq3PQch4hEXSqJY7uZ7X4628xOAnZmLqTGTV1VIhJ1qYxx3A48Z2afEywdewjBUrKRFHdopswhIhGWygOAs82sL9AnLFri7pWZDavx0lxVIhJ1++2qMrObgdbuvsDdFwBtzOwHmQ+tcdIYh4hEXSpjHDeFKwACEC6sdFPGImrkNMYhIlGXSuLISlzEKVwStnnmQmrcNFeViERdKoPj04BnzOyRcPu7wMuZC6lx01xVIhJ1qSSOnwGjge+F2/MJ7qyKHPdgyi51VYlIlKWyAmAceB9YQbAWx1nAosyG1TjFw6ke1VUlIlFWa4vDzI4Grgx/1gPPALj7mQcntMYnrhaHiMg+WxyLCVoXF7j7UHf/E1CVzsXNbJiZLTGzYjMbU8P+08xsrpnFzGxk0r5rzGxp+HNNDedOMbMF6cRzoKoTh8Y4RCTK9pU4LgXWAK+b2aNmdjbBk+MpCe++eggYDvQDrjSzfkmHrQSuBZ5KOrcTcCcwmKB77M6E9ccxs0uBbRxkrq4qEZHaE4e7v+Duo4C+wOsEU490NrNxZnZuCtceBBS7+zJ3rwAmASOS3mOFu88H4knnngfMcPey8LmRGcAwADNrA/wY+E0qFaxP6qoSEUltcHy7uz8Vrj3eHfiA4E6r/ekGrErYLgnLUrGvc+8Gfg/scz0QMxttZkVmVlRaWpri2+6bBsdFRNJcc9zdN7r7eHc/O1MB7YuZnQAc6e6T93dsGGehuxcWFBTUy/t/OcZRL5cTEWmS0kocaVoN9EjY7h6WHci5JwOFZrYCeAs42sxmHnCkKfKwQ00tDhGJskwmjtlAbzPrZWbNgVHAlBTPnQ6ca2Ydw0Hxc4Hp7j7O3Q91957AUOATdz8jA7HXSGMcIiIZTBzuHgNuIUgCi4Bn3X2hmY01s4sAzGygmZUAlwOPmNnC8NwygrGM2eHP2LCsQe1OHMocIhJhqUw5UmfuPhWYmlR2R8Lr2QTdUDWdOwGYsI9rrwCOq5dAU1Q9OK7nOEQkyjLZVfWVo7mqRESUONKi23FFRJQ40qLBcRERJY60aK4qEREljrRorioRESWOtKirSkREiSMtGhwXEVHiSIvmqhIRUeJIy5fPcShziEh0KXGkoUqTHIqIKHGkQ4PjIiJKHGnRJIciIkocadFzHCIiShxpUVeViIgSR1r0HIeIiBJHWvQch4hIhhOHmQ0zsyVmVmxmY2rYf5qZzTWzmJmNTNp3jZktDX+uCctamdk/zWyxmS00s3syGX+iLeWVrNtSDqjFISLRlrEVAM0sC3gIOAcoAWab2RR3/zjhsJXAtcBPks7tBNwJFAIOzDGzKcAu4Hfu/nq4jvmrZjbc3V/OVD2qDb3nNbaUxwAlDhGJtky2OAYBxe6+zN0rgEnAiMQD3H2Fu88H4knnngfMcPcyd98IzACGufsOd389PLcCmEstS8/Wt+qkARocF5Foy2Ti6AasStguCcvq5Vwz6wBcCLxa0wXMbLSZFZlZUWlpaaoxp0TrcYhIlDXJwXEzywaeBh5092U1HePu49290N0LCwoK6vX91eIQkSjLZOJYDfRI2O4eltXHueOBpe5+/4EEmKp49X24IT05LiJRlsnEMRvobWa9woHsUcCUFM+dDpxrZh3NrCNwbliGmf0GaA/cXv8h12zTzso9tpU3RCTKMpY43D0G3ELwhb8IeNbdF5rZWDO7CMDMBppZCXA58IiZLQzPLQPuJkg+s4Gx7l5mZt2BXwD9gLlmNs/MbsxUHaqVbd+1x7bGOEQkyjJ2Oy6Au08FpiaV3ZHweja13BXl7hOACUllJcBB/9besK1ij23djisiUdYkB8cPtg3bkxNHAwUiItIIKHGkYO/EocwhItGlxJGCsqSuKuUNEYkyJY4UJA+Oq8UhIlGmxJECdVWJiHxJiSMF23fF9tjW4LiIRJkSRwq2V1Ttsa3nOEQkypQ4UrCjQi0OEZFqShwp2JHU4tAYh4hEmRJHCnbsqqJ9y5zd20ocIhJlShwp2F4R2yNxKG+ISJQpceyHu7Ozoop2Lb+c1kvTqotIlClx7EdFVZxY3JO6qhowIBGRBqbEsR87dgUD4xrjEBEJKHHsx47KIHG0y9UYh4gIKHHs147wqXG1OEREAhlNHGY2zMyWmFmxmY2pYf9pZjbXzGJmNjJp3zVmtjT8uSah/CQz+yi85oOW4ce4q5/haKfEISICZDBxmFkW8BAwnGCp1yvNrF/SYSuBa4Gnks7tBNwJDAYGAXeGa48DjANuAnqHP8MyVAUguBUXkhNHJt9RRKRxy2SLYxBQ7O7L3L0CmASMSDzA3Ve4+3wgnnTuecAMdy9z943ADGCYmXUF2rn7e+7uwBPAxRmsQ42D45qrSkSiLJOJoxuwKmG7JCw7kHO7ha/3e00zG21mRWZWVFpamnLQyaoHx3U7rohI4Cs7OO7u49290N0LCwoK6nwdDY6LiOwpk4ljNdAjYbt7WHYg564OX9flmnVSPaW6EoeISCCTiWM20NvMeplZc2AUMCXFc6cD55pZx3BQ/FxguruvAbaY2ZDwbqrvAC9mIvhqO6sHx3O/nHJEeUNEoixjicPdY8AtBElgEfCsuy80s7FmdhGAmQ00sxLgcuARM1sYnlsG3E2QfGYDY8MygB8AfwGKgU+BlzNVBwhaHM2zmtGyedbuMrU4RCTKsvd/SN25+1RgalLZHQmvZ7Nn11PicROACTWUFwHH1W+ktduxK0bL5llkN/syx2pwXESi7Cs7OF4fFqzezKI1W2ndPIucrC+zhVocIhJlGW1xNHV3TlnInM820qp51h7PbihviEiUKXHsw7hvn8i905bQrWPLPcr1AKCIRJkSxz50bpfL76/o39BhiIg0KhrjEBGRtChxiIhIWpQ4REQkLUocIiKSFiUOERFJixKHiIikRYlDRETSosQhIiJpUeIQEZG0KHGIiEhalDhERCQtShwiIpKWjCYOMxtmZkvMrNjMxtSwv4WZPRPuf9/Meoblzc3sv83sIzP70MzOSDjnyrB8vplNM7P8TNZBRET2lLHEYWZZwEPAcKAfcKWZ9Us67AZgo7sfBfwRuDcsvwnA3b8GnAP83syamVk28ABwprsfD8wnWJ5WREQOkky2OAYBxe6+zN0rgEnAiKRjRgB/DV8/D5xtwWIX/YDXANx9HbAJKAQs/GkdHtcO+DyDdRARkSSZXI+jG7AqYbsEGFzbMe4eM7PNQB7wIXCRmT0N9ABOAnq4+ywz+z7wEbAdWArcXNObm9loYDTAYYcdVi8VmjR6CCUbd9bLtUREmqrGOjg+gSDRFAH3A+8AVWaWA3wfGAAcStBV9fOaLuDu49290N0LCwoK6iWoIUfkMfKk7vVyLRGRpiqTLY7VBK2Fat3DspqOKQnHL9oDG9zdgR9VH2Rm7wCfACcAuPunYfmzwF6D7iIikjmZbHHMBnqbWS8zaw6MAqYkHTMFuCZ8PRJ4zd3dzFqZWWsAMzsHiLn7xwSJpp+ZVTchzgEWZbAOIiKSJGMtjnDM4hZgOpAFTHD3hWY2Fihy9ynAY8CTZlYMlBEkF4DOwHQzixMki6vDa35uZncBb5hZJfAZcG2m6iAiInuzoFfoq62wsNCLiooaOgwRkSbFzOa4e2FyeWMdHBcRkUZKiUNERNKixCEiImlR4hARkbREYnDczEoJ7sCqi3xgfT2G05BUl8ZJdWmcvip1OZB6HO7uez1BHYnEcSDMrKimuwqaItWlcVJdGqevSl0yUQ91VYmISFqUOEREJC1KHPs3vqEDqEeqS+OkujROX5W61Hs9NMYhIiJpUYtDRETSosQhIiJpUeLYBzMbZmZLzKzYzJrUuh9mtsLMPjKzeWZWFJZ1MrMZZrY0/LNjQ8dZGzObYGbrzGxBQlmN8VvgwfBzmm9mJzZc5HuqpR6/NrPV4Wczz8zOT9j387AeS8zsvIaJumZm1sPMXjezj81soZndFpY3xc+ltro0uc/GzHLNbJaZfRjW5a6wvJeZvR/G/Ey4vAVm1iLcLg7390z7Td1dPzX8EEwF/ylwBNCcYDnbfg0dVxrxrwDyk8r+CxgTvh4D3NvQce4j/tOAE4EF+4sfOB94mWA9+iHA+w0d/37q8WvgJzUc2y/8d9YC6BX++8tq6DokxNcVODF83ZZgcbV+TfRzqa0uTe6zCf9+24Svc4D3w7/vZ4FRYfnDwPfD1z8AHg5fjwKeSfc91eKo3SCg2N2XuXsFMAkY0cAxHagRwF/D138FLm64UPbN3d8gWKMlUW3xjwCe8MB7QAcz63pQAt2PWupRmxHAJHff5e7LgWKCf4eNgruvcfe54eutBIuodaNpfi611aU2jfazCf9+t4WbOeGPA2cBz4flyZ9L9ef1PHC2mVk676nEUbtuwKqE7RL2/Q+rsXHgX2Y2x8xGh2Vd3H1N+Hot0KVhQquz2uJvip/VLWH3zYSELsMmU4+we2MAwW+3TfpzSaoLNMHPxsyyzGwesA6YQdAi2uTusfCQxHh31yXcvxnIS+f9lDi+uoa6+4nAcOBmMzstcacH7dQmey92E49/HHAkcAKwBvh9g0aTJjNrA/wNuN3dtyTua2qfSw11aZKfjbtXufsJQHeCllDfTL6fEkftVgM9Era7h2VNgruvDv9cB0wm+Mf0RXVXQfjnuoaLsE5qi79JfVbu/kX4Hz0OPMqXXR6Nvh5mlkPwRTvR3f8eFjfJz6WmujTlzwbA3TcBrwMnE3QNVi8Pnhjv7rqE+9sDG9J5HyWO2s0Geod3JjQnGESa0sAxpcTMWptZ2+rXwLnAAoL4rwkPuwZ4sWEirLPa4p8CfCe8i2cIsDmh66TRSernv4Tgs4GgHqPCu156Ab2BWQc7vtqE/eCPAYvc/Q8Ju5rc51JbXZriZ2NmBWbWIXzdEjiHYMzmdWBkeFjy51L9eY0EXgtbiqlr6DsCGvMPwV0hnxD0F/6ioeNJI+4jCO4A+RBYWB07QT/mq8BS4BWgU0PHuo86PE3QVVBJ0D97Q23xE9xV8lD4OX0EFDZ0/Pupx5NhnPPD/8RdE47/RViPJcDwho4/qS5DCbqh5gPzwp/zm+jnUltdmtxnAxwPfBDGvAC4Iyw/giC5FQPPAS3C8txwuzjcf0S676kpR0REJC3qqhIRkbQocYiISFqUOEREJC1KHCIikhYlDhERSYsSh0gazOyd8M+eZvater72f9b0XiKNjW7HFakDMzuDYBbVC9I4J9u/nDuopv3b3L1NPYQnklFqcYikwcyqZyG9B/h6uGbDj8JJ5u4zs9nhBHnfDY8/w8zeNLMpwMdh2Qvh5JMLqyegNLN7gJbh9SYmvlf45PV9ZrbAgjVWvplw7Zlm9ryZLTazienOcipSF9n7P0REajCGhBZHmAA2u/tAM2sBvG1m/wqPPRE4zoPpuAGud/eycHqI2Wb2N3cfY2a3eDBRXbJLCSbd6w/kh+e8Ee4bABwLfA68DZwKvFXflRVJpBaHSP04l2BepnkE03PnEcxnBDArIWkA/NDMPgTeI5hsrjf7NhR42oPJ974A/g0MTLh2iQeT8s0DetZDXUT2SS0OkfphwK3uPn2PwmAsZHvS9n8AJ7v7DjObSTB3UF3tSnhdhf5Py0GgFodI3WwlWHK02nTg++FU3ZjZ0eHMxMnaAxvDpNGXYInPapXV5yd5E/hmOI5SQLAcbaOYmVWiSb+diNTNfKAq7HJ6HHiAoJtobjhAXUrNS/NOA75nZosIZll9L2HfeGC+mc11928nlE8mWF/hQ4IZXX/q7mvDxCNy0Ol2XBERSYu6qkREJC1KHCIikhYlDhERSYsSh4iIpEWJQ0RE0qLEISIiaVHiEBGRtPx/Q02l3D+5RggAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(history_data1.history['accuracy'], label = \"train_accuracy\")\n",
    "plt.xlabel('iteration')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.2. Model 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x19cff4672e0>"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(history_data2.history['loss'], label = \"train_loss\")\n",
    "plt.xlabel('iteration')\n",
    "plt.ylabel('Loss')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x19cff4def40>"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(history_data2.history['accuracy'], label = \"train_accuracy\")\n",
    "plt.xlabel('iteration')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 4. SAVE MODELS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_1.save('C://Users/MINH LE/Desktop/TraditionalvsCNN/models/Experiment 1/Exp1-MNIST-TraditionalFilter.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_2.save('C://Users/MINH LE/Desktop/TraditionalvsCNN/models/Experiment 2/Exp2-MNIST-TraditionalFilter.h5')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}